Meta

Transcript: (disclaimer: may contain unintentionally confusing, inaccurate and/or amusing transcription errors)

Ben: I was up late last night, late for me as a dad is like 11:00 PM, but I’m sitting here at my computer in my dark basement, pulling notes together. David, what music did I have on? You have one guess.

David: Oh, I know exactly what you had on. You had Trent Reznor’s The Social Network soundtrack.

Ben: It makes anything you’re doing feel twice as important and twice as revolutionary. It just felt very apt for this episode.

David: Oh man. I have been listening in the last 24 hours to 50 Cent In Da Club because that came out my freshman year of college, same year as the facebook.com. Man, 50 Cent, facebook.com, can’t get any better than that.

Ben: Perfect match made in heaven. Well, I’ll check out our wall to wall from the old days and see if there were already posts about that. All right, let’s do it.

David: Let’s do it.

Ben: Welcome to the Fall 2024 season of Acquired, the podcast about great companies, and the stories and playbooks behind them. I’m Ben Gilbert.

David: I’m David Rosenthal.

Ben: And we are your hosts. Today, we are studying a company whose products are used by more humans than any other company in history—Meta—of course, formerly known as Facebook. I figured I would contextualize these numbers a little bit. Meta has four billion monthly active users.

David: And daily active users are over three billion.

Ben: Yeah, nuts. There are only eight billion humans on earth. As I started to brainstorm what the closest competitors could be to serving half of the humans, I thought, surely I can find it in empires or governments from the past where there is some larger percentage.

David: Yeah, it makes sense.

Ben: Nope. The Roman Empire at its peak was only 40% of humans, tops. The data’s a little bit hard to actually find from that period of time. The British Empire, which we have a little bit better handle on, at its peak was only 23% of the global human population. No government, tech company, utility, et cetera has ever addressed so much of the world.

David: It’s just wild. There’s no other way to put it.

Ben: In the over 20 years since its founding, Meta truly has connected humanity through its apps—Facebook, Instagram, WhatsApp, Messenger, and now Threads. Today we’re going to study how they did it.

There’s been a lot of inks spilled writing about Facebook over the years, and for its first 10 years, most of that writing focused on the many benefits to society, with breathless exuberance over milestone after milestone.

For the past 10 years, it’s seemed like Meta could do nothing right. Reporting focused on its many stumbles, the massive mistakes, the incredible controversies surrounding the company. While we will of course discuss these events as part of our story and analysis, our goal today on this podcast episode is really to understand how it is that Meta became the dominant fabric that connects the human race, and why they’ve been so successful at continuing to win over and over again.

No matter what you think of the company, it is undeniably one of the most important institutions in the world, and their global scale is no accident. It is the result of careful actions from some of the most motivated and brilliant people in the world who believe in one mission—connecting as many people as possible.

David: I hadn’t thought about this till now, but almost taught logically this is the most important episode we’ll ever do, to try and understand this. It’s the biggest company that has ever existed. It’s the biggest thing that’s ever existed.

Ben: Yeah, by these measures that we’re talking about right now. Now, of course the story of how we got here is nuts. It is the perfect Acquired stew like you’re saying, David. They speed ran their startup phase, they swerved through multiple disruptive technology waves, they battled fierce competitors, they invented or maybe discovered one of the greatest business models ever, and they’re now trying to pull forward the next technology generation through sheer force of will with AR, VR and AI.

Finally, listeners, we tackle one of the greatest corporate stories ever: Facebook, the Mark Zuckerberg production.

Listeners, we have one big announcement for you today. It is time for our annual Acquired survey. If you have 3–5 minutes, please click the link in the show notes or go to acquired.fm/survey to take it. We’ll be raffling off a pair of shiny new Meta Ray-Bans and giving away a bunch of ACQ dad hats as well.

This is really our one big ask of you all year and it helps us immensely with making the show better, to hear your suggestions, and also to help our sponsors understand just how impactful the Acquired audience is. Go to acquired.fm/survey and David and I are both eternally grateful.

After this episode discuss it in the slack, and check out ACQ2, our second show where we just had Clem Delangue, the CEO of Hugging Face on, to talk about his view on how the open source AI ecosystem will play out.

Before we dive in, we want to briefly thank our presenting partner, JP Morgan Payments.

David: Just like how we say every company has a story, every company’s story is powered by payments and JP Morgan payments is part of so many companies’ journeys from seed to IPO and beyond.

Ben: With that, this show is not investment advice. David and I may have investments in the companies we discuss, and this show is for informational and entertainment purposes only. David, Roman Empire, what’s our starting place here?

David: No man, we got to start in 1984. We’re never going to finish this episode. Really, where else can you start? I can make up Roman Empire Classics, Escalus, whatever T-shirt saying Mark has this day of the week, but Facebook has always started and ended with Mark Zuckerberg.

We start in May of 1984 when Mark Elliot Zuckerberg is born as the second of four children to Karen and Ed Zuckerberg in Dobbs Ferry, New York, small little suburban town, suburb of New York City.

Ben: His mom is a psychologist and his dad is a dentist, is that right?

David: That’s right. Ed studied dentistry at NYU. He was always really into math and computers, though. But growing up the son of Jewish immigrants in New York City, the expectation was you’re going to be a doctor or a lawyer. He goes to dentistry school, becomes a dentist, he meets Karen, they get married.

After Ed finishes dental school, they move up to Dobbs Ferry in Westchester County. Mark and his three sisters grow up as prototypical middle class suburban American kids in the nineties. Life is good. Your grandparents worked really hard to make a bunch of sacrifices for your parents. Your parents are now professionals. They’re also working hard to give opportunity to you, and the family is leveling up in America. It’s the classic story here.

Ben: You can play sports, you can hang out with your friends, and play video games.

David: Indeed. For Mark and for me growing up and maybe you too, it’s amazing. There’s no reason not to just explore your passions. For Mark, those passions become threefold: (1) turn-based strategy video games, most particularly Sid Meier’s Civilization, (2) programming computers, and (3) as you mentioned and everybody who watched our Chase Center show knows, ancient Greek and Latin history.

Ben: Classics.

David: The classics, and pretty much in that order of importance both to him personally, and the future course of Facebook and Meta. Lt’s start with Civilization. September, 1991.

Ben: David, I toyed with the intro to this show being, “And we are your hosts. Civilization is a video game upon which…,” but I decided not to.

David: It’s laughable, but no, this is really important here. September 1991, when Mark is seven years old, Civilization or Civ as its fans call it, comes out for PCs, and young Mark gets hooked.

For those of you who haven’t played Civ, it is a very particular type of turn-based strategy game, and is probably the foremost example of what is known as the 4X genre. 4X stands for explore, expand, exploit, exterminate.

Despite that nefarious-sounding nature to the last one there, especially, Civ in 4X games were this radical branch of video games at the time where the action of playing and the way to win wasn’t just about killing all the enemies on screen. It’s a lot more like a board game or like Risk if you’re familiar with that. It’s about strategy, it’s about growing and marshaling your resources effectively. Most importantly, there are multiple ways to win the game.

Most other games out there you can win by defeating all the other players. But you can also win by completing a technology development tree, and reaching the peak of technology, which is in the game, launching a spaceship to go colonize a new planet. That’s another way to win.

In modern versions of the game, you can also win by diplomacy. By getting elected the leader of the UN. You can also win by dominating all the other players on either culture or religious dimensions too. It’s like this amazing simulation of what it’s like to run one of the top 10 market cap mega capitalization companies in the world today.

Ben: But to your point, this idea that it’s turn-based, you try to amass resources, figure out how to deploy your resources, have multiple concurrent strategies so that as the roll of the dice of the universe happens and things unfold before you…

David: And other players make their moves, yup.

Ben: Right, to be able to effectively react to it all and use your resources to win in whatever way winning means to you.

David: It’s funny. This is way too simplistic of an analogy and not fair to Mark on any dimension, but the whole construct here reminds me of the book Ender’s Game, where it’s these kids growing up playing this game and they didn’t realize all along it was actually the real game.

Playing Civ and other video games, when he is young PC games, leads young Mark to want to learn how to program himself. When he’s 10, he has his parents take him to the local Barnes and Noble one day and begs them to buy him C++ For Dummies. He dives into the book, learns everything in there, and he’s like, I need more. I need more.

After that, his parents hire a local tutor to come and tutor him for an hour a week in programming and learning how to write software. Mark has this quote that he says later, “I’d go to school and I’d go to class and come home. The way I’d think about it was awesome. I have five whole hours to just sit and play on my computer and write software. Then Friday afternoon would come along and it would be like, okay, wow, now I have two whole days to sit and write software. This is amazing.”

Ben: It really is astonishing as a 10-year-old to be learning C++. This is not BASIC. This is a language where you are thinking very close to the metal. You have to be very aware of the constraints of your system, of manually managing memory. It’s impressive for 14 or 16 year olds to be learning simplistic languages. This is a whole different ball game

David: As it always is with Mark. While he’s a kid here in Dobbs Ferry, he codes up all sorts of little projects including a family chat tool for all the networked computers in the house. Also his dad’s dentistry office is attached to the house, so he builds this network called ZuckNet, which is family chat between all the computers, very prescient here. Then, speaking of family and friends chat, in 1997 when Mark’s 13 years old, AOL launches AIM (AOL Instant Messenger).

Ben: Oh, I remember it well. Many, many, many hours lost to AIM, warning my friends and other dynamics of the AIM system, to gang up on people and to have other people gang up on me.

David: Oh man, punting people off AIM.

Ben: Having to create a new screen name, log off and log on, message someone, and pretend you’re someone else. This was middle school.

David: This was middle school for all of us. This was middle school for Mark, too. The crazy thing about AIM, just a quick little digression here, I had no idea till doing the research, it was actually a rogue skunk works project within AOL. Did you know this?

Ben: No.

David: It was a couple of engineers that just went off and built this almost against the wishes of management. It was not originally intended to really be for social use or for kids at all. It was more like the idea they had was it was going to be like Slack, like chat for workplaces. They ended up using it internally at AOL to communicate, and that’s actually where away messages come from.

They came up with away messages because people would be in meetings. This is the days of desktops. There are no cell phones, barely even laptops. People would go out to lunch and they’d have to put up a message so that when their colleagues were trying to communicate with them, be like, no, I’m in a meeting. No, I’m away. That’s the origin of away messages.

Ben: That’s fascinating. It is interesting how you ended up using away messages after a while, even while you were there just to indicate status.

David: It was the ultimate emo teenager thing. It was like I am expressing my whole self through this one line away message.

Ben: 100%.

David: For people all of our age, Mark’s age at the time, that completely changed our social experience when that came out. The other important thing about AIM when it came out that I had no real conception of at all as a kid was that it was free. My family was an AOL family. We were AOL subscribers, we had the, You Got Mail, et cetera, but you didn’t have to be a paying AOL subscriber to download and use AIM and join the network.

AIM goes on to become this total viral thing in North America. Kind of concurrent with Napster a little bit before Napster. It’s the first piece of consumer viral software. By 2001, a couple of years after launch, it has 36 million users, including, as we’ve been talking about, basically every middle- and high-schooler with access to a computer.

Thirty-six million users doesn’t sound like a lot today, but you got to remember the Internet was way smaller back then. Way, way, way smaller. Not just in terms of number of people had access, but also number of people who were literate in how to use it.

Ben: Totally. It’s the early broadband era is that era we’re talking about. You’re right. Tens of millions of people, at least in the US, had internet access at this point in time.

David: Yup, and basically all of them and certainly all the young people were on AIM. The other important thing about AIM and AOL here is it was pretty easy to hack. Oh man, there was just a field day that kids and teenagers had with AIM bots, punting software, warez, hacking programs.

If you were someone like me, you were just downloading this stuff off of FTP servers on the Internet, installing it on your machine, and it let you be a punk and kick people off of AM by bombarding them with messages. You could do some teenage punk stuff. But it also let you extend the functionality of what AOL and AIM was. You could add graphics, you could add emojis, you could draw pictures.

Ben: Add links in your profile. Oh the ASCII art in your profile, in your away message.

David: Totally. This becomes such an important part of teenage culture on the Internet as we’re growing up is all this personalization of, hey, I don’t want just what AOL gives me. I want to control this. For me, I was downloading and using this stuff. You might have been writing it, although you might have been a little young.

Ben: No. I didn’t learn PHP until I was 15 or 16, so call it 2005.

David: Okay. So yeah, a little later.

Ben: Yup.

David: Mark is one of the kids that’s writing this stuff. I think that’s where he got his start. He was doing ZuckNet, he was doing all these little projects, but here’s AOL, this unintentionally, infinitely hackable and extensible canvas that he can play around on.

That’s number two for Mark, passion number two, computers and programming. And then passion number three is classics. Mark, when he enters high school—he’d gone to public school all through elementary and middle school there in Dobbs Ferry in Westchester County—he does two years at the local public high school at Ardsley High School.

After two years at Ardsley, Mark is like, well a couple of things. I’m not sure that this is the right high school for me for the rest of my high school career. (1) I think I would enjoy just being in a better school, a higher level of competition, higher level of learning. He’s definitely at the top of the class at Ardsley. (2) I’m really into Latin and classics and I’m going to tap out on that real fast here at Ardsley. There’s got to be a place I can go where I can really study this. That takes us to Phillips Exeter Academy.

Ben: In New Hampshire, right?

David: Yes, in New Hampshire. Exeter is one of these private boarding high schools that have existed basically as long as Ivy League colleges have. They’re there just as feeder schools into Ivy League colleges.

Mark applies to Exeter, gets in, and the summer before he’s about to go, the school hosts an event in New York City, for incoming two years to meet each other, get to know the school, get ready, et cetera. There, Mark meets another kid who’s a lot like himself, who also is really into computers, really into AIM and everything going on there. Mark’s like, yeah man, I really made the right choice. Everybody here is going to be just like the two of us. That person is Adam D’Angelo, future CTO of Facebook.

Ben: Amazing they met in high school.

David: Now, the great irony of this story is that Mark and Adam ended up being the only two kids who are really into computers in the class. It was just the two of them. They just happened to meet in New York City at the preschool event.

Ben: Which is probably one of the last few years that that would be the case. I’m sure if you went to this school today, it’d be 80% of the classes or 100% of the class is really into computers.

David: I think in large part because of Mark and Adam, it’s quite different these days.

So the two of them, when they get to Exeter become fast friends. They work on all sorts of computer projects together, software projects, and by the end of their senior year, to graduate from Exeter you have to do a senior project. Most people do a mini college thesis or they do something in the community, et cetera. Mark and Adam are like, no, we’re going to build some software. This is what we do.

Mark one day is listening to music on his computer. I think probably on Winamp, which is how everybody listened to MP3s that you got from Napster back in the day.

Ben: I think I listened on SoundJam but it’s just because I had a Mac.

David: Oh yeah, future Seeds of iTunes.

Ben: That’s right. It got acquired by Apple, and then eventually they (I think) built off the same code base and turned it into iTunes. But everyone used Winamp. That was the AIM of music players.

David: And in fact, AOL I think would acquire Winamp a couple of years later, bring it all into the fold. Legend has it, Mark is listening to a playlist in Winamp one day during his senior year, and he gets to the end of the playlist.

He’s like, why did the music stop? I listen to so much music all the time on Winamp, it should just know what I like and automatically pick the next song. It should just be a DJ. Then inspiration hits, he calls up Adam, and he is like, I’ve got our senior project. We are going to build an AI DJ plugin for Winamp, and they call this Synapse.

They did it. These two high school kids. They build an AI, AI wasn’t called AI at the time, but a recommender system that looks at the data of what songs you’ve consumed on Winamp, and predicts what song you’re going to like to consume next. It’s basically Spotify today.

Ben: What year is this?

David: This would’ve been spring of 2002.

Ben: Real early days.

David: So they build this thing and then they release it to the Internet. People start using it. It starts getting some adoption, and a couple of companies including Winamp itself reach out to Mark and Adam and are like, hey, do you want to come work here? Can we buy this?

Ben: I think Microsoft was one of them, too. It was for like a million dollar acquisition offer or something that wasn’t crazy. But these are high school kids.

David: In 2002. But yeah, this is really important because they’re about to graduate. Adam is going to go off to college at Caltech. Mark is going to go off to college at Harvard in the fall. But Mark is not just any student coming into Harvard. He’s like, I can build these things and they have value, and big companies are going to be willing to pay me for them. I don’t need to be that worried about getting a job.

Ben: He always knew he had a great fallback plan, so it let him be risk on.

David: So Mark graduates, he goes off to Harvard, it’s the natural choice on many fronts. Mark, of course, aces his SATs. I think he gets a 1600 on the SATs. There’s this great legend about you had to take SAT twos to get into college back then. I don’t know if you still do, but I think you had to have a minimum of three.

Mark doesn’t study or prepare for them at all. He walks in the day of the test, says I’m going to take all three that I’m going to take on the same day. He gets eight hundreds on all of them. College admissions was not going to be a big deal. And his older sister, Randi, was already there at Harvard.

Ben: So clearly go to Harvard if he wants.

David: Yes. I think he wants enough. I don’t know that it’s that important to him, but he is like, yeah sure. Harvard, it’s the best I’m going to go there.

Ben: Yup. In fact there’s a video of that.

David: That’s right.

Ben: Have you seen this? His dad filming his reaction and he’s pretty measured about it. It’s very, yup, got into Harvard, and then goes back to doing something else on his computer.

David: Yup. I think for most kids, for most people in the world, you would look at that and be like, what arrogance? And look, it’s not like 18-year-old Mark is not arrogant, but I think it’s also, this is why all this stuff we’ve been talking about is important. It’s just not that important to him.

He likes building stuff, he likes making stuff. He knows it has value, he knows people will use it. Whether he goes to Harvard, doesn’t go to Harvard, goes to college, doesn’t go to college, it’s just not that big a deal to him I don’t think.

Ben: Reading in between the lines, it seems like the thing that he was most excited about is oh good. A challenging environment where I will encounter other really smart people like myself who are ambitious.

David: I think that is totally right. Mark gets to Harvard in fall of 2002 and like you said, he goes to class, he does well, but he wants to meet other people and be around other smart folks. He ends up joining the AEPi fraternity there, like my co-host here, my esteemed colleague Ben.

Ben: Indeed.

Ben: Also my father while he was in college.

Ben: I didn’t know that.

David: That’s right, AEPi. Adam meanwhile of course goes off to Caltech, 3000 miles away in the opposite corner of the country.

Ben: Also, can I just say, how crazy is it that Adam D’Angelo would become Facebook’s first CTO? They don’t go to college together.

David: Well, I was going to get into this in a sec. Adam has a very different experience at Caltech. He ends up graduating from Caltech. I think they probably do not have the same generous leave policy there that Harvard does.

Ben: Because Harvard’s is weird, right? It’s like you can leave and if you ever want to come back, it’s like you never left.

David: We’re going to come back to that in one sec. Put a pin in that. But for Adam at Caltech, he still finds time while he’s there to keep working on these projects.

During his freshman year, he builds something called BuddyZoo, building on Synapse and the success that they had in their senior project. BuddyZoo takes your buddy list from AIM, from AOL, and effectively turns it into a social network. You upload your entire buddy list to BuddyZoo, and then you get your friends to do so too.

BuddyZoo then analyzes and gives you a list of who are the friends of your friends, how many degrees of separation do you have to them It automatically identifies cliques, like who of your friends are part of these cliques and who are their friends. It measures popularity among users. This is starting to sound pretty familiar here.

Ben: This is pretty advanced stuff for 2002–2003.

David: Very advanced stuff. Especially for a side project for a very busy Caltech student. BuddyZoo, pretty quickly in the months after Adam launches it, gets a couple of hundred thousand users. There’s gravity to this thing. It’s becoming fairly big, especially given how small the Internet was at the time.

Adam is like, whoa, okay, this is interesting. This is what scale looks like. Synapse was cool and we had a bunch of people downloading it, but this feels like something more on the order of AIM itself. This is really interesting.

Ben: And the important thing with Synapse was it was a single user application, so it could analyze my songs and tell me what to listen to. This is the first time Adam, Mark—particularly Adam in this case—is discovering a very different type of application that gets better as more people join it.

David: Now, we say Adam and Adam definitely built and launched it, but it’s not like Mark wasn’t involved. They’re staying in touch the whole time collaborating and working through coding challenges together on what else? On AIM. It’s serving its original use case. They’re collaborating through it across the country.

That summer 2003, after their freshman year, Adam comes to Boston and gets an internship at the MIT media lab. He discovers this thing called Friendster, another social network out there. Adam is like, whoa, this is a dedicated standalone version of what I just built with BuddyZoo. This is really interesting

Ben: And it’s all on the web

David: Now, Friendster was based out here in Silicon Valley, had raised money from Kleiner Perkins.

Ben: And Benchmark, right?

David: Oh, Benchmark too? I didn’t know that. Interesting.

Ben: Yup, which is what tied their hands and why they could not pursue an investment in Facebook.

David: That’s funny.

Ben: Not that I think that was a real possibility anyway because things were moving so fast. But you never hear them in any of the discussions about who was pursuing Facebook and who was iced out. They just had a competitive investment.

David: That’s right. It’s funny, I never even think about that because obviously Matt Kohler would go on to become an important partner at Benchmark.

Ben: And of course Matt was an early executive at Facebook.

David: Exactly. Interesting. So here’s Friendster, this example of a real standalone company, social network, growing virally, funded by venture capitalists, based in the San Francisco Bay Area. Adam’s like, oh, maybe that’s what we should be focusing on.

By the end of the summer in 2003, Friendster had over three million users. Really interesting growing really quickly.

Ben: Friendster had real scale and grew really fast. That is something that has been lost to history. Friendster is the butt of a lot of jokes, but it exploded out of the gates.

David: True. Adam’s playing around with Friendster, and Friendster had this feature called testimonials, which was like a Facebook wall. But in the long run, one of the core problems with Friendster was it positioned itself as a social network and about friends, but really it was intended to be a dating site underneath it all.

Ben: Absolutely. That was wink-wink not a dating site. It was this idea that there are all these other things that are dedicated to dating. There’s this weird stink about it. But if you just naturally meet a friend of a friend online through a thing that’s totally not a dating site, then great. There’s no stigma around it.

David: So specifically when it comes to testimonials, the purpose of testimonials was supposed to be, why you should date somebody or why they would be a great partner, which is funny.

Adam, while he’s playing around with this and he’s telling Mark about it, he actually posts on Mark’s Friendster profile a testimonial and he says that “Mark gets way too lucky.” He posted this in the summer of 2003.

Adam was telling me this story and he was like, you know? I have thought about that post for 20 years since then, because on the one hand it was like a stupid throwaway, double entendre that I put on there because it was a dating site.

On the other hand, it is both absolutely true about Mark, that he always gets way too lucky and Facebook and Meta always gets way too lucky. But it’s not luck. It’s actually Mark, there’s something very different about the way he operates, and all the reasons that Meta and Facebook get lucky over the years, Mark was always that way, and it’s always because of Mark.

Ben: Mark is unbelievably good at positioning himself for luck to do its thing.

David: And specifically at this point in time, what Adam talked about is he was always looking around. He was looking at Friendster, he was looking at what we did with Synapse. He was looking at what I did with BuddyZoo. He was looking about the projects at Harvard he did that we’re about to talk about. He was always thinking, how do I incorporate all of this and make whatever I’m doing better?

This was the stew that Facebook is about to come out of. This brings us to fall semester 2003–2004, Mark’s sophomore year, and the famous Kirkland House Suite H33 with Mark Zuckerberg, Chris Hughes, Dustin Moskovitz, and Billy Olson as the four roommates there.

Ben: Three of the four of those names are names you may know as the founders of Facebook.

David: Yes. Mark is focused heading into his sophomore year at Harvard. He is not focused on classes, he’s focused on projects.

Ben: And he’s done like 10. This is something that again is lost to history. People talk about, oh there’s the one FaceMash thing that he did that led to Facebook. He did like 10 side projects.

I remember being in this era of my life where you feel like you have superpowers as a programmer, you’re looking around and you’re like, oh I can make a website for that. I can make a website for that. I had the same thing where the world wasn’t saturated with apps yet, so you could just make things that made your life better or other people’s lives better or cool ways to connect people.

I distinctly remember feeling, how come nobody else realizes you can just do this? I know what that feels like and I know that mentality he was in. There were 10 things that he had worked on even before FaceMash.

David: Yup. The first of which, right as they’re getting back to campus that fall, is Mark builds a tool called Course Match.

Ben: Oh yeah, that’s right.

David: And this is, as students are shopping for classes for fall semester, you can upload which classes you’re thinking of taking or which classes you’ve signed up for, and you can see who of your friends have also signed up for that class or are planning to sign up for that class. This is just text. All this is is a list of classes and a list of people who sign up, and it takes Harvard by storm.

Think about it. It’s obvious in hindsight. A bunch of college undergraduates, yeah you want to take classes you’re interested, but really you want to make sure you take the classes with the people that you want to be around.

Ben: I was literally texting my friend yesterday who was at Harvard at this time. She said, oh yeah, people totally chose their classes based on who was in them.

David: Of course. Mark sees this and he’s like, again it’s one of these learnings, an observation about the world. He’s like, wow, people are spending a lot of time on this tool.

Ben: It’s literally just lists of people. You click on a class name and people would spend a lot of time just combing over that list of people.

David: Totally. There’s nothing visually compelling here. He’s like, well okay now it’s time for my next project. People have registered for classes. What if there was something visually compelling for people to spend time on?

Ben: What if I overstepped a little bit and…?

David: And this is FaceMash, well-chronicled in all the many books, stories.

Ben: The Stephen Levy book, the David Kirkpatrick book.

David: Movies, what have you. Mark codes this up in one evening in his dorm in Kirkland House, and it’s basically the website Hot or Not for Harvard, with a couple of actually interesting mechanical twists on it.

Now, the two big problems with it are one, this is probably just not a good idea or not kind, Generally, it is a head-to-head voting mechanic on which picture is “hotter” than the other picture. The other thing that is really not great is rather than having users upload their own pictures, mark hacks into the Kirkland House servers and downloads the face book photos of all the students to populate the website.

Ben: Lowercase face space, lowercase book.

David: Yes. We should probably spend a minute and talk about what a face book was before Facebook. My big question for you is did Ohio State have a face book?

Ben: It did not.

David: It probably would’ve been a telephone book–size thing if they did.

Ben: I went to college in the fall of 2007, and you could look up people’s email addresses if you had their first and last name. You might be able to get some other maybe major, but it was a text blob. It wasn’t a photo.

David: Well, I think most, if not all Ivy League schools back in the day had these. Princeton, where I went, certainly had these when I showed up on campus.

Ben: In fact, Phillips Exeter had one. The private high schools did, too.

David: That’s right. You had to give the university, or in Exeter’s case the high school, a school photo of you when you were coming in before you matriculated. Then they would literally make a book, a printed book with the names and class years and photos of all the incoming students from the freshman class.

I think if I remember right, at Princeton at least, they would republish the book every year for your class because some people would leave, some people would join, et cetera.

Ben: All right, so these are photos stored digitally that Mark then, while plugged into the campus network from his dorm room, is accessing, downloading, and then putting up on his own website that he’s hosting.

David: Yes, called FaceMash. Now, a couple of interesting things about FaceMash. One, gosh, if Mark was blown away by the engagement of the Harvard population with Course Match, FaceMash takes that to such a whole nother level that it actually melts down the servers in Kirkland House and the IT Department at Harvard has to shut off internet within hours of it going live to all of Kirkland House, lest this continue to spread virally within Harvard and take down all of the servers.

Ben: Oh, so maybe he didn’t actually rehost the images. Maybe he was just pointing at the URLs that were hosted by Kirkland House.

David: Oh, I didn’t think about that. Yeah, that might have been how it worked.

Ben: But either way, dude knows how to make something that gets engagement.

David: Yes, he does. The other interesting thing from Mark’s perspective about FaceMash is that it was the first time he was building a rich web app as one of these projects, fully built on the LAMP stack. And he had probably used the LAMP stack (Linux, Apache MySQL, PHP) for Course Match.

Ben: To be clear, that’s an operating system, a web server, with Apache MySQL, the free open source database, PHP, the free open source programming language.

David: Yes. But this is the first time that he’s building a real rich consumer-grade web app on top of the LAMP stack.

Ben: Rich consumer grade. It’s the technical requirements of Course Match but with photos.

David: Photos and voting.

Ben: Sure. Yeah, you probably have an additional table of information in the database or something. But the big takeaway here is: (a) of course the use case shouldn’t have done it, (b) wow, kicks engagement, (c) this is training wheels of how to use these new open source web technologies.

If you go back two, three, four years, you’re going to have to go to Oracle, you’re going to have to go to Microsoft, you’re going to have to buy enterprise-grade super proprietary systems to do this. This is all just free, something you could cobble together, upload onto some Linux web hosting, and boom. There had to be a hundred times more, a thousand times more web applications created by the Mark Zuckerbergs everywhere in this period of time doing whatever random little project that they thought would be fun or funny or useful.

David: It’s such a perfect example despite its problematic nature of this is the first time in history where you need neither money nor permission to launch an application like this on the Internet.

Ben: Totally. $10 domain, $100 a year web hosting that includes the database, no licenses required. PHP is all free. Maybe the bandwidth would’ve been an issue but this is on the order of $100 to do this.

David: Well, clearly the bandwidth was an issue because they shut down the Internet to Kirkland House.

Ben: So David, this feels like a thing you should get kicked out of school for doing.

David: Yeah. Harvard as you imagine is not pleased about this. They haul Mark in front of the administrative board, which is a disciplinary committee for students when things like this happen. He and his fraternity brothers are pretty sure that he’s actually going to get thrown out of school for this.

The night before the ad board meets to make their final decision AEPi, the fraternity, throws a goodbye Mark party that night. Legend has it that this is where Mark and Priscilla Chan meet for the first time. Priscilla, of course, is now Mark’s wife. Dad board meets the next day and decides not to throw Mark out of school but to put him on disciplinary probation for the rest of the year. Basically, don’t do this again.

Ben: Slap on the wrist.

David: Slap on the wrist. So Mark doesn’t end up getting kicked out of Harvard. But he does end up becoming the Harvard computer celebrity even before Facebook.

Ben: This is a big deal. Everyone on campus knew about Facemash. A lot of people already knew about Course Match anyway. He’s now the guy who can make websites and web applications here at Harvard that people use.

David: And it just so happens that there is a group of upperclassmen at Harvard led by two twins, the Winklevoss twins who have an idea to digitize Harvard’s face book and build an online version of it that they’re thinking of calling the HarvardConnection. Then they think maybe we could actually take that to other schools too and call it ConnectU as the broader version. This was not at all a unique idea.

Ben: Not at all.

David: Lots of people at lots of schools had this exact same idea. In fact, Harvard itself had this idea. The IT Department at Harvard had been promising for years and talking about how they were going to build a digital version of the Harvard face book.

Ben: And Friendsters out there. And I looked it up. Stanford had Club Nexus, Columbia had CU Community, Yale had Yale Station. I’m sorry, there’s a whole movie made about the drama of how novel this idea is. It’s not.

David: Totally not. I’m pretty sure Princeton had one too when I was there. Pre-Facebook. Yeah, totally not a novel idea. Lots of people have it.

Ben: Nope.

David: The Winklevoss twins though, Mark is now this celebrity on campus, they need somebody to be a programmer to write the site, so they talk to Mark and Mark agrees that he’s going to help them out and help code this site that they’re thinking about.

Ben: Also, again, the chutzpah to say we have this idea, you are a programmer, you will program our idea, clearly Mark can come up with his own ideas that get people excited enough to use the stuff he builds. Now, so far it’s been a little unsavory, but he doesn’t have a problem coming up with ideas with product/market fit and executing them end to end. The only thing he hasn’t done so far is made money, made anything as a business. But what other pieces of the puzzle does he really need from some people with an idea?

David: None. It’s funny, I was going to save this for a little later, but let’s talk about it now. In addition to technically because of open source and the LAMP stack, and also socially because of AIM and everything going on, the world for the first time, something like Facebook being able to be built for no money and no permission.

Also, this was the first time that a technical person, because of this, could just do everything. Mark didn’t need anybody. Mark didn’t need a non-technical co-founder. Mark didn’t need a CEO. He was the CEO.

Ben: I remember thinking the same thing when I shipped my first app to the app store. Co-founder Ian and I in 2009 (I think) made something called Seize the Day. Got over a million downloads, two programmers uploaded something to the app store. That’s a slightly different era because that’s mobile, not web, but I remember looking around being like, whoa, we didn’t need a business guy. Right. That’s the craziest thing. You can make stuff and you can put it in the world.

David: Totally. This takes us to Harvard’s Christmas break. Now at the time Harvard did final exams for the fall semester after Christmas break in January. The idea is all the students go home for Christmas break, you come back in January, and then there’s a week of reading period where you re-familiarize yourself with all the material from your classes. Then you take final exams, then there’s a little break, there’s a course shopping period for spring semester, and spring semester starts in February.

Ben: So Mark has a few weeks to program a new idea.

David: Exactly, to not study and program a new idea. During Christmas break, he had actually come out to Silicon Valley, to the Bay Area to hang out with some friends out here, and seen the physical embodiment of all these great tech companies—Yahoo, early Google, et cetera, drove by them all. He comes back and he’s extra, extra motivated. He’s like, you know? This digital face book idea, I think this is something that a lot of people want. I bet I can code this up in a week and just launch it.

Ben: And again, not a business, just a project. He has this gut feeling that people will use this if he makes it.

David: Totally. Now importantly though, all the stuff he’d been doing throughout the year, he’s like, well I can incorporate that into this project too. It’d be pretty boring if it’s just a digital face book. I don’t think people want that, but I’ve made all this other stuff that people really, really like and engage with.

Ben: And when you say just a digital face book, you mean essentially the profile page would just be a photo and a name.

David: Right, just like the physical version of here is a list of all the students in the class and here is their stock profile picture.

Later, after the facebook.com blew up, Mark told the Harvard Crimson in an interview. “I don’t really know what the next big thing is because I don’t spend my time making big things. I spend time making small things, and then when the time comes I put them together. That’s the stuff I do. Small, little projects, and eventually they all fit together.” That’s what the facebook.com ends up being.

Mark gets back from reading period. Instead of studying, he dives in. He IMs his friend, Andrew McCollum, who he knows from CS classes, and asks Andrew to do the page design. Andrew says, yeah, sure thing, I’ll help you out. Andrew goes and fires up Photoshop. He finds an image on the Internet of some guy who looks cool and puts it in the header of the design for the site.

Ben: Behind a cloud of ones and zeros because it’s cool.

David: It’s like the Matrix.

Ben: It’s gradiented, it’s white on the left, blue on the right.

David: And for a year or two, however long the Facebook guy was at the header, everybody’s like, who is the Facebook guy? People think it’s Al Pacino. People are like, it’s a student. Do you know who the Facebook guy was?

Ben: I totally do. I looked this up. I spent two hours trying to figure this out.

David: I did too.

Ben: Really?

David: Yeah, it’s amazing.

Ben: Because everyone always said it’s Al Pacino and I’m like, it doesn’t really.. Al Pacino must have looked really different when he was younger and it’s not.

David: All right, tell me your answer. I’ll tell you if it matches mine.

Ben: All right, so there was an 80s song that got really big called Centerfold.

David: Oh yes. This is it.

Ben: And the image is of Peter Wolf.

David: Yup, who is the lead singer of the J. Geils band.

Ben: We’ll link to that image. There’s this great Quora post, which we should say Adam D’Angelo founder and CEO of Quora after Facebook. There’s a great Quora post about this where there’s unmistakably the photo of Peter Wolf that Facebook guy is based on. They just pixelated it and made it duo tone instead of the original photograph.

David: So great. On the evening of February 4th, 2004 after final exams, right when students are coming back, getting into the mode for the new semester, starting to shop for spring semester courses, Mark Zuckerberg launches the facebook.com. And right off the bat from the moment somebody logs in, the functionality is awesome.

The registration page reads, you can use the Facebook to search for people at your school, find out who are in your classes (Course Match), look up your friend’s friends (BuddyZoo), see a visualization of your entire social network. BuddyZoo plus graphics. Then at the bottom of every page is the famous copyright 2004, the Facebook, a Mark Zuckerberg production.

Ben: Awesome. I feel like it’s like one of the most infamous screenshots is the screenshot of that original homepage.

David: So, so great.

Ben: Importantly, things that did not exist here yet. Messages, wall posts. Even I don’t think pokes were in the original very first version.

David: That’s a good question. If they weren’t in the very first version, they launched pretty quickly thereafter.

Ben: But no Facebook events, no photos other than your profile photo. I don’t even think status updates were in the very first one.

David: I don’t think status updates were there yet, no. But all the lessons from all the other projects, all the elements of them are coming together here. Once again, registration is limited to email addresses on the harvard.edu domain.

That was part of the magic of what had made all the other projects so successful, is like it’s not randos in here. It’s other people at Harvard. It’s your classmates, it’s the people you care most about. And by the nature of that, it’s their real identities

Ben: And that’s what makes it different from all these other social networks. MySpace existed, Friendster existed, but anyone could sign up for these. And in part, that meant that they had more explosive growth because there was no governor on the growth.

We’ll talk about all the problems that come from anyone anywhere can sign up at any time. But the very core thing here is authenticated people. The people who were signing up for Facebook at the start are people you know go to the same college as you, and have to use their real name matched against a university-issued email address. Identity and authentic identity is a part of the company from its first moment.

David: Coupled with super important learning number two from FaceMash, it’s all user-submitted content. No pre-population with anything from Harvard, no scraping from Harvard servers, et cetera.

On the one hand that prevents Mark from getting into trouble. On the other hand, though, it sets the expectation and the requirement and trains the behaviors of the users of you got to enter all the content into this site yourself. You put in your AIM screen name, you put in your cell phone, you put in your interests, you put in your classes.

Ben: Also, isn’t it crazy that they displayed cell phone numbers for the longest time?

David: Well again, think about the functionality and the utility to the users. On the one hand, you would never disclose your cell phone number to the general public. On the other hand, other students and your friends at Harvard knowing your cell phone number, well that’s actually pretty useful for getting in touch, for going out, for going to parties, for planning things, et cetera.

Ben: These are all people you would give your email and phone number and maybe even birthday too if they just asked.

David: Totally. You would probably also give them your relationship status, which is in there. Formalizing your relationship status for the first time here on the facebook.com. You can add a photo of your own choosing as your profile picture. So big difference from old school face books where it’s whatever the university chooses or it’s a stock school photo or whatever. No, you get to express your personality. You get to Photoshop it as much as you want.

Ben: And you can update it whenever you want.

David: And you can update it whenever you want.

Ben: It is amazing how much of the next billion-plus people’s source of value comes from this founding moment. The Facebook grows tremendously in functionality over time, but everything is like a natural outcropping out of authentic identity, user submitted content, trust that the people who are seeing this content are people in your private network. The whole thing is here in this first few weeks of coding that got done and then threw up the landing page.

David: I would even go one step further than that. The actual nature of the seeding of the network here is maybe the most important thing. Here’s what I mean by that.

Once Mark launches this, it takes off like wildfire within Harvard. Of course, he’s already a celebrity, already known for launching these projects. People are tuned in waiting for the next Mark Zuckerberg production.

Within the first 24 hours, 650 people join. There are only 2000 people in a class at Harvard. I think there are about 7000 or 8000 undergraduates total. Within two weeks, over half of the entire undergraduate population is active on Facebook.

Now, here’s what I mean and why the seeding is so important. This initial network is super dense, super engaged, and super active. That not only sets the norms for how you’re going to behave on Facebook going forward. It also means that as the network grows virally, it radiates out from this nuclear reactor of a base.

Contrast that with an approach of something like Friendster that’s default open to everybody, backed by Silicon Valley Venture dollars, trying to grow, they’re doing marketing. Okay, let’s say somebody in Ohio hears about Friendster, somehow maybe they see an ad for it, they sign up. When they join the service, there are no friends there. There’s nothing to do. There’s no functionality. It’s not alive.

When you join the Facebook, even though it’s limited to just Harvard students, you sign up and you’re blown away by how alive this site is.

Ben: Almost nobody can use the Facebook, but for the people who can, it’s an unbelievably great experience. As they expanded, that stayed true. It was either you’re not allowed to use it or if you are, it’s blown away great, instantly.

David: That’s basically always true. The moments when Facebook stops growing is when that’s not true, and then they refocus on it, and then that gets the growth back.

Ben: Yup.

David: Here’s another thing that I think was really important about this seeding of Facebook, was actually that it happened at Harvard. Harvard is the top global brand among universities. Whatever you personally think about Harvard, it’s undeniable, especially back then that it was the top.

If you were a kid or a family pretty much anywhere in the world, you knew what Harvard was. This was really important as Facebook expanded because it had this patina to it. It was elite, especially as they grew. There were clones, there were lots of other Facebook copies out there.

Ben: And even if they were as good as what Mark had built, they didn’t start at Harvard.

David: It’s like Civilization. Nobody else could ever invade Facebook’s turf, but they could invade other people’s turf. Somebody put it to me in the research, there was a fairly big competitor in Germany, I think maybe called Study Vice or something like that. Facebook could really easily go into Germany and people in Germany would be like, oh Facebook, I’ve heard of that. This is interesting. Study Vice could never go into Boston, ever.

Ben: It’s interesting. Yeah, that’s a great point. Okay, so weeks go by and they’ve saturated Harvard. Not only that, you start seeing this thing happen, which just to level set with listeners, this stayed true for several years. This is a crazy stat. 70% of people who ever signed up were active that day.

There are not user engagement retention metrics better than that in the world. The fact that as they grew, it didn’t matter how many more people they added, it was still the case that 70% of users ever were daily active. So you have this crazy situation at Harvard where it’s just like, to your point, nuclear reactor levels engaging.

In fact, this term wouldn’t get coined till later in the summer. But this is from The Facebook Effect, David Kirkpatrick’s book. They had coined a term for how students seem to use a site. Zuckerberg, Moskovitz, and Parker, Sean Parker, who we’ll talk about who comes in, called it the trance. Once you start combing through the Facebook, it was very easy to keep going. It was hypnotic. You just kept clicking and clicking and clicking from profile to profile, viewing the data.

David: And every single one of those clicks is a page view. Yeah, this hits Harvard like an earthquake. It’s FaceMash all over again, except a good problem, not a bad problem. It’s all legit. The servers are melting down. Even Mark is blown away by how big this thing gets so quickly.

Ben: Right away they add the wall. There’s a way to publicly post things like testimonials on your friend’s walls. You could view a wall to wall so you could see the public posts that you were making back and forth to each other. You’ve got your profile information, wall posts, and that’s it.

David: But to your point, very quickly Mark is adding features, adding functionality, adding things to do to the site. The other thing that he does right off the bat is say, okay, we’re taking this to other colleges too.

He recruits his roommates, Dustin Moskovitz comes on to help maintain the site. Dustin was actually an econ major. He had no idea how to code. He goes out and buys Perl for Dummies, and Mark’s like that’s great that you read that. The site is written in PHP. Dustin goes and learns PHP too and helps with coding the site and keeping up with everything.

Ben: You know Perl, you can learn PHP. It’s much easier.

David: Chris Hughes joins too. Andrew McCollum joins, and then they need to keep buying servers to keep up with the traffic or else everything is going to melt down.

One of Mark’s fraternity brothers in AEPi is a guy named Eduardo Saverin, and Eduardo was (I think) part of the business and investing club at Harvard. They recruited Eduardo in, and Eduardo invests $1000 in infrastructure for the site. Mark invests another $1000. Together there’s $2000 of investment in the facebook.com.

The deal they strike is that Mark gets two-thirds of the company and Eduardo gets one third of the company. Eduardo sets up a business entity to formalize all of this as a Florida LLC. Eduardo was from Florida.

Right away, as we were saying, Mark’s like, great. This is not just a Harvard thing. We need to bring it to other colleges. He specifically chooses Columbia because Ben, as you said, there was an existing competitor there and he’s like, I want to see how Facebook competes. Like, I want to know right away can we displace it?

Ben: It’s so interesting. He doesn’t go to the schools where there’s nothing. He goes to the schools where there is something and wants to be better than it.

David: Which is so counterintuitive to how most people would think. They say, oh no, this is a race. I’m going to go to the white space. I’m going to gobble everything up, and then wait to take on competitors when I’m bigger and stronger. Mark’s like, nope, right away.

Ben: This is an early mindset of we are going to be globally dominant. If you’re okay splitting the market, you go for the white space and then you say, let’s see how much of the low hanging fruit we can easily get. But if you’re Facebook, you’re going to say, let’s go to the hardest possible competitor, extinguish them, and then we’ll have a better shot at owning the whole market and we can get to the low hanging fruit later.

By the way, their strategy once they did win at a school like Columbia, was to go figure out all the schools that were closest to Columbia in terms of network connections, and then win there too, to build a moat around their victory.

David: There we have the initial crew, Mark, Dustin, Chris, Andrew, and Eduardo, the five co-founders of Facebook. The other interesting thing about going to Columbia, which they do three weeks after the launch at Harvard—22 days, it was the same month all within February, 2004—they set up Columbia as a completely different network than Harvard. If you join the Columbia Facebook, you can’t then go access all the profiles of Harvard students.

Ben: At first there was zero connectivity between the two.

David: Which is another counterintuitive thing. You’d think like, oh great, I want to build this as big as possible. Mark clearly wants to conquer all these other schools. The best way to do that is to have all this live content come right to them from Harvard. I think he realized that had he done that, it would’ve socially inhibited sharing within Harvard.

Ben: The scarce commodity is trust. That is the important lesson here. It’s also a very convenient infrastructure decision, where if these two systems truly don’t need to talk to each other, that’s great. Put different servers in different data centers. Don’t worry about overloading the database. Don’t worry about number of writes per second and reads per second. It’s great. If every school gets its own server rack, you end up not having some of these issues that the other social networks have.

Friendster’s out there trying to compute second degree friends of friends, which is this crazy hard computer science problem, especially as the global number of people growing. It’s an N-squared problem. Facebook is over here (or the Facebook at the time) in the land where they’re keeping N small. Even anything that’s N-squared is contained within that school, and their servers aren’t falling over and taking 20+ seconds to load pages the way that the Friendsters of the world are.

David: This is such an interesting point because technology infrastructure advantage has always been a core part of Facebook, even back to this very beginning. What’s interesting is that it flips at some point as Facebook grows and eventually becomes open to everybody. Their competitive advantage becomes their tech is so good and their infra is so good that they can do everything you were just talking about that Friendster couldn’t do and that MySpace couldn’t do. But in these early days, their competitive advantage is actually like, no, we’re scale out, not scale up. Then Mark was totally willing to just flip the bit later.

Ben: It’s effectively counter positioning. They’re building the Facebook in such a way that the use cases don’t require incredibly sophisticated technology to accomplish those feature sets. There’s this funny chicken or the egg thing here and I think the answer like many chicken or the egg is both. Mark, at the moment he conceives of an idea, thinks through the technical requirements and the user experience. They’re co-mingled in the product development process. I really do think even at age 19, he was aware of the scaling benefits in addition to the user experience benefits of launching in this way.

David: Totally agree. Okay, fun bit of trivia about Harvard’s leave policy that we referenced earlier. Do you know what else Mark does on the same evening that they launched Columbia?

Ben: No idea.

David: He goes to hear Bill Gates speak on campus.

Ben: That was the same night?

David: The same night during which speech Bill mentions that he actually felt comfortable dropping out of Harvard because he discovered that Harvard had this really generous leave policy that you could take an infinite leave of absence and pursue something else and come back anytime you want. Mark would actually say later that Bill saying that at the event was part of what got him comfortable doing the same thing.

Ben: Unbelievable. It is crazy. Tying this back to our Microsoft episode, this would’ve been what, early 2004? This would’ve been right after Bill stopped being CEO, handed the CEO reigns to Steve. He was still the chief architect and still chairman of the board. But it’s this post-DOJ time for Microsoft where Bill is just technically focused and can do things like go speak at Harvard.

David: It is amazing how much influence Microsoft had on Facebook and on Mark. We have a lot more to talk about on that front.

Next couple of days after Columbia launches, Facebook launches at Stanford, at Yale. By the end of the semester, they’re in over 100 schools in 3–4 months. The speed they’re moving at is crazy.

Ben: And they start building these wait lists. There are so many common startup things that Facebook invented. They built tremendous demand before they would light up the network. They knew that as soon as you’re in Facebook, you want to quickly get to 7 or 10 friends or whatever the metric is to create that magic moment where you’re like, oh yeah, Facebook now works for me.

They wanted to wait until they had sufficient demand to boom open that school, and then once it’s open everybody should have the best possible experience. You’re seeing Facebook basically say, okay, wherever there’s really strong demand, that’s where we’ll open next, and we’re not going to open anywhere where we see middling demand for our product.

David: Meanwhile, here we are at the end of the school year. The Winklevoss twins take the whole ConnectU situation to Harvard’s president, Larry Summers, to adjudicate this dispute between them and Mark. Larry says Harvard is not going to get involved here. This is a business dispute between students.

Eventually the two twins sue Mark and Facebook. They settle for $65 million, $45 million of which is paid in pre-IPO Facebook stock. Obviously, that grows. A large portion of that then I think they sell and put into Bitcoin in 2011–2012 timeframe. They end up pretty fine out of all this.

Ben: They did great.

David: They did great. But here we are at the end of the school year, summer comes around, and Andrew McCollum has some connections out in the Bay area. He had interned at Electronic Arts. Adam D’Angelo is going to be spending the summer from Caltech.

The crew all decides, hey great, let’s not get internships. Let’s go rent a house out in Palo Alto, move to Silicon Valley, and work on Facebook there for the summer and see what happens. Yes, the house did indeed have a pool. Yes, they did indeed install a zip line off the chimney. That was not fiction in the movie. That did actually happen.

Ben: And it is important to know to this point, it is a project. Mark, Dustin, Chris, and everyone who moves out, are there to work on a project that they think is cool, seems to be working, and get exposure to Silicon Valley and venture capitalists for when they start their startup having a network. That is literally the mentality. Facebook is live at how many schools, David?

David: A hundred.

Ben: A hundred schools, and they are going out not with the intent to make this a company, but to contemplate what company they could start and meet people that can help them with that.

David: Now when we say crew who moves out, that crew did not include Eduardo. Eduardo had an internship in New York and decided that he was not going to move out for the summer.

Ben: Yeah, tough decision.

David: Not a great decision. On the other hand, he ends up with 2% of Facebook, so he’s also fun.

Ben: To make a long story short, yes, everything you saw in The Social Network around this, the result is effectively correct that Eduardo Saverin goes from owning a third of this Florida-based LLC to something like 2% of this C-corp that is a Delaware C-corp based in California, that goes on to become or is Facebook Inc.

The justification that they effectively use in changing this structure is, hey, a bunch of us moved to California to start a company together. You stayed back and yes, you sold some ads in the meantime, but you didn’t come start the company with us. Listener, we leave it to you to decide how that should have played out and what’s fair. None of us were there. He was a co-founder of Facebook and then ended up with 2% and lived on a different coast.

Okay, so they’re in California. They get really serious about the Facebook. They actually are still working on something else called Wirehog concurrently within the same team. But they’re starting to realize, okay, the Facebook thing really, really has legs. As they’re contemplating their next move, they literally run into someone on the street who will change everything. David, who is this person?

David: That person is the co-founder of Napster, Sean Parker, who would have a brief but very, very large impact on the company.

Ben: Before we tell the Sean Parker chapter of the Facebook, now is a great time to talk about our presenting partner JP Morgan Payments. In these critical moments like we are talking about now, where a new paradigm or technology shift happens, David and I typically focus on the founders and the innovators who figure out how to build something really great. But unfortunately it’s not just the good actors who take advantage of these opportunities.

David: There are always bad actors on the other side of the same new technology. This is especially true whenever money or value is exchanged, whether it be phone scammers taking advantage of international networks in the 80s and 90s, viruses spreading through the early internet, or today’s cybercrime like AI-powered identity theft. As technology gets smarter, so will the fraud tactics.

Ben: Enter JP Morgan Payments. With the digital payments ecosystem expanding rapidly through innovations like real-time payments, blockchain, and peer-to-peer networks, the attack surface has grown exponentially. For a company like JP Morgan who moves $10 trillion a day across 160 countries, staying ahead of bad actors is critical for the entire global financial system.

That’s why they’ve been investing in technology like AI and cyber fraud prevention for years. Their AI models have been trained on billions of transactions. That is unparalleled scale and expertise when it comes to fraud intelligence, and it’s only getting smarter and more reliable given their processing over 5000 transactions per second with 99.9% accuracy. Regardless of your size or industry, you can leverage their enterprise-grade solutions like validation services, which validates bank accounts, individuals, and businesses.

David: One great example is a Fortune 500 global media company. They needed an automated solution for the thousands of influencers on their platform to do rapid onboarding, minimize errors, and ensure consistent validation of bank account information. They were able to compress validation times from days to minutes so they could scale.

Ben: They’ve got a great payments developer portal. It’s a trusted, scalable platform built with payment, security, and risk management in mind.

David: Developers can even integrate their preventative fraud solutions to act before an incident takes place. Sardine, a leader in financial crime prevention collaborates with JP Morgan Payments and embeds its account validation services API to analyze and protect over 2 billion devices for 300 companies with AI to the highest standards of fraud and compliance safety.

Ben: David, I actually met the founder of Sardine at the meetup that we did after the Chase Center event with JP Morgan Payments in San Francisco last month.

David: That’s right.

Ben: It was awesome. Ultimately, every business benefits from built-in fraud prevention. Whether you’re moving $100 or $1 billion. And with the trust and innovation from JP Morgan Payment’s whole ecosystem, you can drive growth while protecting your business. Check out more payment solutions and stories at jpmorgan.com/acquired.

Okay, David, Sean Parker enters the picture. Here we go.

David: Sean Parker had of course started Napster with Shawn Fanning back in the late 1990s. After the whole Napster saga, Parker ended up starting the email contact list company called Plaxo, which was its own proto social network and pioneered email address, book exploitation, and exporting.

Ben: Which is how every social company got its start. Facebook did a ton of this. LinkedIn did a ton of this. WhatsApp did this.

David: Exactly. It pioneered the first bootstrap viral social network mechanic.

Ben: Hey, let me see who all the people you know are, so that on this new thing that you want to be connected to all the people you know, you’re connected to all the people you know. Oh by the way, we might also invite all the people that you’ve ever contacted.

David: Yup. Sean and his co-founders at Plaxo raise money for Plaxo from venture capitalists. And not just any venture capitalists. The best venture capitalists from Sequoia Capital. All is going well. Paxos going fine. Sir Michael Moritz is on the board of the company.

Anyway, after a couple of years, Sean Parker ends up getting kicked out of the company by the board. Now Sean wasn’t exactly the most reliable employee and definitely was quirky, but the result of this is Sean’s everlasting enmity for Sequoia, paranoid distrust of all venture capitalists and company boards. This is going to become very, very, very important here in a sec.

Rewind slightly back to the spring of 2004 when the team is still at Harvard. Sean is living in Palo Alto, and one of his housemate’s girlfriends who’s I believe a student at Stanford pulls up the facebook.com on her laptop. It’s like the scene where this happens in The Social Network movie. Sean is like, holy crap, this is it. This is the winning social network. Sean, I believe had actually been an advisor to Friendster as well. He was between Plaxo and Napster. Friendster really believed social networking was going to become a thing.

Ben: Totally. It is crazy to think about, unless you had a student show you on their computer this is what Facebook is, there’s no real way for you to know other than the signup page.

David: Which you can’t get past it and see how alive the thing is. Sean, being the enterprising person that he is, cold emails the company’s business email address and asked to set up a meeting. Eduardo Saverin replies to Sean like, oh cool, you’re the co-founder of Napster. Yeah.

They set up a dinner in New York City. This is while the spring semester is still going on at Harvard. Mark, Eduardo, and Sean all get dinner together, and Mark and Sean really bond. Mark is super pumped to meet him. But that’s about it. There’s no discussion about what’s cool, a billion dollars, et cetera, as dramatized in the movie. That doesn’t happen. They think that they’re just going to go their separate ways.

Ben: Nice to meet you. I don’t think even Mark knows yet that he’s moving to California this summer.

David: No, totally not. Here we are now in June, the Facebook crew has moved out to Palo Alto. They’re walking down the street one evening. Sean is also walking down the same street in Palo Alto, sees them and is like, hey, it’s the Facebook guys.

At this point, Sean Parker was living in a different house, specifically the house of his girlfriend’s parents—not a good situation—and he makes a split-second decision right there on the street to insert himself into Facebook and is like, hey, I’m actually looking for a place to stay. Do you think I can crash with you guys in the crash pad for the summer? He joins the crew and this is how it all starts.

Ben: That is wild.

David: Yeah. It’s like, have you ever played old school Super Nintendo role-playing games, like Japanese RPGs, like Final Fantasy or whatnot?

Ben: It’s been a while.

David: It’s just like you meet somebody on the street and they join the party. You always know the character on the street that’s going to join the party because their pixel Sprite is more highly detailed than the regular NPC pix. This is exactly what it is like here.

Ben: Sean Parker had a very detailed pixel spray.

David: So while he’s there living in the house crashing with the college students, he basically takes Mark aside and he is like, look, you’ve got something really special here. I have been in your shoes. Let me tell you exactly what is going to happen this summer.

You have created something magical. You’re now out here in Silicon Valley. You are going to be the bell of the ball. All the venture capitalists out here, all the big companies, they’re going to wine and dine you. They’re going to court you, they’re going to tell you you’re great, and they’re going to want to invest in your company.

Then they’re going to turn around and they are going to screw you. They’re going to take control of your board. They’re going to force you to bring in “professional management.” They’re going to kill all the magic, growth is going to slow, then they’re going to blame you, then they’re going to fire you, and then they’re going to put the company up for sale.

Ben: And here’s the thing. To this point in history, he really wasn’t wrong.

David: He was absolutely right.

Ben: There wasn’t Founders Fund yet. There wasn’t A16Z yet. There wasn’t the notion of founder friendliness. What venture capitalists did is they invested in founders companies, and then brought in management to take them to the next level.

It’s what happened at eBay. It’s what happened at Google. This is a pattern that happened over and over and over again. Usually, they would try to do it with the help and support of the entrepreneur to build this really great team together. But oftentimes, that backfired and they just hot swap the CEO with a big grownup manager in a nice suit.

David: That was not the wrong playbook until this moment in time because as we’ve been talking about, until now if you wanted to scale something, you needed money, you needed permission, you needed business people, you needed deals, you needed contacts.

Ben: That’s true. The open web changed the hard requirement on that.

David: And I think Sean, as bitter as he was about his own experience, he probably also recognized that, and recognized that the Facebook and Mark had the potential to start a new paradigm of how to do this differently.

Sean says to Mark, look, if you just go out on your own, that’s what’s going to happen. I am going to take you under my wing and make sure that that does not happen. I’m going to help you reset up the company so that you permanently control the board and no one can ever fire you because you are what is important here.

I’m going to help you go about fundraising, but I’m going to make sure we find the right people who are cool with that stipulation number one, because if you go try and do it by yourself, you don’t know enough, you don’t know these people, you’re not going to find the right folks.

It’s wild. Sean was also in a position to take advantage of Mark and the company, and he didn’t.

Ben: There was some nice self-interest in all of this. It’s not like this was charitable.

David: He ends up being the president of Facebook.

Ben: And he ends up with a pretty decent chunk for a non founder, and he ends up with a board seat. Now granted it’s technically Mark’s board seat, but Sean’s the one sitting in it.

David: But it would’ve been really easy for him to do one of two things: (1) Obviously take advantage of Mark and the company and get a lot more for himself. (2) I think he also could have been so jaded by the Plaxo experience to say, we shouldn’t raise venture capital at all.

But actually what he counsels Mark to do is the right thing. No, we should raise venture capital. We should professionalize and the whole Silicon Valley machinery really can help us, but we’re going to do it in the right way on our terms. I think he just really wanted to help Mark win the game.

Ben: I think you’re right.

David: The other thing that Sean does by all accounts is he makes sure that everyone in the house that summer gets equity in the company.

Ben: Okay. What’s Wirehog? What’s the second project that’s going on here?

David: I really would love to know what Sean thought when Mark and the crew told them about Wirehog. Wirehog was basically going to be a captive Napster to Facebook. Everybody on Facebook in these private network’s walled gardens, what is something that college students really, really wanted to do and was happening internally on college service all the time? Share files, share music, share movies. Wirehog was going to be a product to make this happen.

Actually legend has it, I think in Steven Levy’s book, while they’re talking about this, Sean Parker actually comes up with a name for the product. It is like, oh, if you launch this, you should actually call it Dropbox.

Ben: No way.

David: Yeah, a year before Dropbox, the actual company gets founded.

Ben: That’s so funny. But they really are thinking of themselves as this almost like an incubator lab. I don’t really understand how this is true, but they really did convince themselves that their current front runner for the product they were most excited about was Wirehog.

Facebook might be a good distribution vehicle for it. Or maybe at some point they don’t even focus on Facebook anymore and they go all in on Wirehog, and they’re actively talking about this insanity while speaking with investors about raising capital.

David: To the investors that Sean goes out and helps them find, Sean at this point knows basically everybody in Silicon Valley and he’s like, okay, I need the pretty tight window of people who: (a) have money, (b) know what they’re doing, and (c) are okay with our terms that Mark is going to control the board and control the company.

Ben: We need an individual, not an institution.

David: He calls two angel investors who he knows pretty well and he thinks might be the right fit. One is Reid Hoffman, who Reid had been running LinkedIn and Sean had been in this early social networking thing.

Ben: And he just started LinkedIn. They actually knew each other because Reid was an early investor in Friendster.

David: Sean calls up Reid and he also calls up Mark Pincus. Mark had been Sean’s boss at an internship that he had in high school. Then Mark had gone on to invest in Napster when they started Napster as a company.

Ben: We should say who Mark Pincus is for anyone who doesn’t know. Mark would eventually start Zynga.

David: Yes, could become a very important person and company in the Facebook ecosystem.

Ben: The maker of Farmville.

David: Yes. So Pincus and Reid meet up with Sean and Mark, and they’re just blown away by the Facebook. Ben, as you say, Mark starts talking about Wirehog and apparently Reid is like, no, no, no, no. Stop talking about this Wirehog thing. I’m running LinkedIn, I can tell you this is special. Do the Facebook thing.

Ben: Not only do the Facebook thing, but music is Kryptonite as a space right now because of what Napster did to the music labels. How is Sean not the one being stay a thousand yards away from music? My God.

David: Amazingly a couple of years later, Sean being so open-minded, was one of the ones who really helped Spotify get going.

Ben: And really brokered the relationship with Mark, for Daniel and Mark to hit it off and obviously Spotify to have a huge amount of growth on Facebook.

David: Totally. So, okay. Read is really excited. He is like, no, no, the Facebook. Focus on the Facebook.

Ben: Your user numbers are what? Your engagement is what? Your retention is what? Your DAU/MAU ratio is what? Oh my God.

David: Reid’s like, look, I’m running LinkedIn, Pinkus and I don’t have that much money together. We’ll give you some, but we can’t fund a whole round for you.

Ben: And they’re conflicted both with Friendster and LinkedIn.

David: But he is like, I think I know the perfect person who can put this together.

Ben: I worked with this guy at PayPal,

David: My old colleague from PayPal, who actually as luck would have it, is thinking about starting his own venture capital fund. This could be a really good fit. Why don’t you meet Peter Thiel? And Peter, yes, indeed was just in the process of starting what would become Founders Fund.

Ben, you already alluded to this, but Founders Fund was such a radical idea when Peter started it, the idea being baked into the name that they will always side with founders and never push a founder out of a company.

Ben: I was shocked. It’s so easy to forget this, but it was just 20 years ago. The idea that a venture capitalist makes an investment in a company and they are not the controlling shareholder was blasphemous just 20 years ago.

David: Totally blasphemous. Of course PayPal has its own crazy history where you can understand why Peter would arrive at this conclusion. But he really was maybe the only person in Silicon Valley that hit that Venn diagram that I talked about a minute ago of super legitimate, would actually help the company and help them navigate Silicon Valley, had money because the PayPal exit had just happened, and was going to be cool with Mark controlling the company.

Ben: It’s a one of one intro to make. They probably would’ve had to be someone else in the PayPal mafia that fits that.

David: But I don’t think at that time, Elon wasn’t doing investing. I don’t think there was anybody else who really could have done it.

Ben: That’s a crazy counterfactual. Thinking about if Elon had invested in Facebook instead of Peter Thiel.

David: What if Elon hadn’t started SpaceX?

Ben: He wasn’t just as good a position. Had Roelof gone to Sequoia yet?

David: If he had, he would’ve just started.

Ben: This is effectively the pool of financiers who could have done this deal.

David: But Sequoia wasn’t going to do this deal under those terms. Roelof couldn’t have done it. Yup. It was maybe one of two with the other one potentially being Elon. Wow, that would be a really different world that we would be living in today.

Okay, so they take the idea to Peter. In the meeting, Reid brings along a young guy who’s working for him at LinkedIn named Matt Cohler to come help talk to Peter, evaluate this from an investing standpoint. They strike a deal that Peter’s going to lead a round of $500,000. Reid and Mark Pincus are each going to invest $37,500. Peter’s going to do the rest at a $5 million pre-money valuation, so $5.5 million post, which on the one hand is laughable now.

On the other hand, at this point in time, for a bunch of kids who started this a couple of months ago, it’s crazy that this would be worth $5 million.

Ben: Out of $5 million valuation, the multiple cents then has been about 250,000X.

David: That takes us to the end of the summer. Everybody now has to make a decision. Do they stay in Palo Alto and keep working full-time on Facebook, or do they go back to school?

Eduardo never came out in the first place. He did his internship in New York. He goes back to Harvard and starts his junior year. Chris Hughes also goes back to Harvard, starts his junior year, graduates in 2006, and then rejoins Facebook after graduation. Andrew McCollum stays in Palo Alto for a couple of years and then goes back to Harvard later and completes his degree. He takes the Bill Gates’ road-not-traveled option. Dustin never goes back, stays with the company until 2008 when he leaves to start Asana. Mark, I think probably never really even considered going back.

Ben: At this point, the 70% stat keeps being true despite the fact that they keep opening all these new schools. It keeps being true that 70% of people who have ever signed up are daily active users. I think Mark realizes, oh, I have created one of the most engaging technology applications ever.

David: And certainly at this point, the combination of Sean, Peter, Reid, Mark Pincus, all advising and being part of the company, they know and they have certainly helped him realize that. Adam D’Angelo does go back to Caltech and then through AIM, I guess, remotely helps collaborate with the company until he graduates in 2006 and becomes CTO officially.

Ben: Got it.

David: Sean, like we said, his impact on the company was huge. I really want to underscore this again. Mark, having full control over the company as we are going to see time and time and time again as we go through the story here, makes all the difference in the world. That is 100% because of Sean Parker.

Ben: How long is Sean at the company?

David: Not very long because at some point in time, the next spring.

Ben: This is nine total months at the company.

David: Yeah, I think about nine total months. Sean ends up leaving the company after charges get filed against him as a result of some drug incidents at a house party during a trip to North Carolina.

Ben: Charges do eventually get dropped, but in the meantime the decision is made, hey, this is not something the company’s going to get sucked into.

David: Yup. He immediately joins where else? Founders Fund, which Peter partially on the back of this Facebook investment, has now set up and raised officially. Like we said, Sean goes on to find and fund and really help nurture Spotify through his role at Founders Fund.

Ben: It’s amazing thinking about Founders Fund raising capital on the back of this investment. One pitch is we invested in Facebook. The flip side of it is we just invested in a company, we’re currently holding it flat because there’s been no markup yet from any other investors, and they’re not generating any revenue.

David: Well let’s talk about other investors. By this point, towards the end of spring semester 2005, Facebook had now dropped the ‘the,’ become just Facebook. That was one of Sean Parker’s final impacts on the company, is he actually negotiated the domain name purchase of facebook.com, the company.

Ben: I didn’t know that.

David: Yeah, I believe they bought it for $200,000, which actually was a good chunk of that venture money that they raised.

Ben: Because it was $500,000 that they raised before.

David: It was $500,000, yup.

Ben: And they’re actively spending that. The company has never consumed that much capital, but they are having to spend it on servers. As they stand up new colleges, they are having to go into data centers and rack servers or rent servers on a monthly basis. Server bills are starting to add up, especially as they scaled to how many schools were they at in the fall of 2004?

David: Several hundred.

Ben: Well, Sean Parker introduced Mark to Western Technology Investment (WTI).

David: The debt fund, yeah.

Ben: They basically invented venture debt and they make a $300,000 loan. I think a revolving credit line in the fall of 2004. That comes with warrants that if the company ever goes public that they could then exercise, they do a deal again the next spring for another $300,000, so they’re in for $600,000. I spoke with someone years ago who told me this ended up being phenomenally, phenomenally successful and the warrant exercise on this I think is the greatest venture debt deal of all time.

David: Has to be. Nice. Spring 2005, Accel comes in, leads a series A, $12.7 million at a $98 million post money valuation. That’s crazy in and of itself for the time, and though still being okay with Mark having board control. Which, one thing at the angel round maybe that’s still pretty crazy but whatever. This is angel round, but no, a real venture capital firm, real series A, 2005, (a) $100 million valuation, (b) being okay with that, that was wild. Oh, also (c) getting less than 20% of the company. Rule of thumb venture investing back in the day was you need at least 20% of the company, ideally more like 25% or 30%.

Ben: Yes. Kevin Efrusy, who ran down the deal and did the diligence at Accel and Jim Breyer who was the partner and took the board seat, this is one of the all time great venture investments, so it’s worth talking about some of the deal mechanics.

It was the very first post dot-com institutional venture capital deal where the founder maintained control. Certainly the norm was as soon as a VC gets involved, it is a VC-controlled board and company.

The other thing that’s worth noting is this is a company getting a $98 million valuation. Now, the silly land that we live in now, we’re like this happens all the time. This didn’t happen and we were only just coming out of the dot-com era. Eyeballs and clicks had just had four years of demonization from everyone, from the press to the public to limited partners.

In fact, Accel had limited partners drop out of this fund who were LPs previously, including your beloved Princeton, including Harvard, I mean big, and I think Stanford was one of the only ones that really stuck with them for this $400 million fund. They’re looking at this realizing this might be one of the greatest companies of all time, and we are going to have to do the type of deal that everyone got raked over the coals for five years ago in the mania for doing. No one is doing this type of deal in this environment, but it’s Facebook so we’re going to do it. I think the level of risk and reputation risk that they took on this cannot be underscored enough.

David: Totally. I was trying to think of, were there any other deals like this, that you could even point to as comparable from the past? The only one I can think of was the Google series A, which was $25 million at a $100 million post money valuation, split between Michael Moritz from Sequoia and John Doerr from Kleiner Perkins.

Ben: But it was in the dot-com runup.

David: Exactly. It was a totally, totally different paradigm. Legend has it—I don’t know if this is apocryphal or not—that after making the investment, Moritz told Don Valentine at Sequoia, never have we paid so much for so little. I think he was actually referring to the small amount of equity that they got by having to split the deal in Google. Yeah, in the dot-com bubble that happened, but that this would happen in 2005, Accel went way out on a limb for this.

Ben: Some other interesting deal points. $1.1 million of this $12.7 million round was done by Jim Breyer personally.

David: Wow.

Ben: The shares were acquired at 4.50 cents per share. Facebook just recently hit $600 a share. Accel and their limited partners and Jim, for anyone who’s still holding the shares after they distributed them, that is a 13,000X return.

Another interesting thing on this, do you know the whole Don Graham dynamic with this deal?

David: Yes.

Ben: The original deal was that Don Graham from the Washington Post was going to invest and it wasn’t going to be a VC deal that was at a $60 million valuation. Accel comes in over the top and in part of the negotiation to get it up to this $98 million deal. There was actually a secondary.

David: Oh wow.

Ben: $3 million went to Mark, Dustin, and Sean as a secondary in this deal.

David: Wow. I had no idea. Wow. Add to the litany of things that were not done back then.

Ben: This is Facebook having all the leverage and full deal control in negotiating this series A. And still, it was one of the best venture capital investments of all time.

David: Absolutely. Man, that is freaking wild. Wow. Okay, well that is a lot of hype to live up to. Which brings us now to the summer of 2005, where Mark, motivated by this perhaps I hadn’t quite thought about that, announces to the company his product roadmap for the summer. It is a six-point product plan.

Number one, a redesign of the site. Bye-bye Facebook guy. Number two, a photo application.

Ben: Which, by the way, photos got written by one guy in two weeks.

David: Yes. Move fast and break things. Number three, a personalized newspaper based on all of your friends’ activity. Number four, an events feature. Number five, a local business product. And number six, a vague idea for a feature called I’m bored, which would let users on the site who were bored consume media and play games.

Ben: That’s a pretty big feature mandate.

David: This is freaking wild. Of that, I think only the redesign and photos would actually ship in 2005. Maybe events did too. I’m not entirely sure, but this is incredible. Photos, newsfeed, platform, which is really what this I’m board feature is. Summer 2005 it’s all right there in the vision, and Mark thinks they can accomplish all of this by the end of the summer.

The wild thing is they do accomplish all of this in the next two years, which two big points on this. For any other company, these are all like multi-year long development processes.

Ben: Well, and each one is its own company.

David: Exactly.

Ben: Photos is Flickr. There are independent game platform companies out there.

David: Miniclip, et cetera. Yeah. Flash Games.

Ben: What were some of the other ones?

David: Events. Eventbrite. Personalized newspaper. Well that becomes a true, true innovation in newsfeed, but I think at the time there was a lot of buzz and talk about personalized portals. Yahoo was really big on this and Google was even big on this of, oh, make this your homepage.

Ben: Oh, iGoogle. Do you remember iGoogle?

David: iGoogle, exactly. I think that might have been somewhat of the inspiration, although obviously Mark was thinking way bigger than that. But yeah, again, local business product, Yelp. Each one of these under the old paradigm was its own company and Mark was like, no, this is all part of Facebook.

Ben: Crazy.

David: Super crazy. This really I think speaks to the genius of Mark as a product strategist. It’s one thing to say, I’m going to add a lot of random features to my site and I’m going to throw spaghetti against the wall. This was not a random grouping of features. Everything reinforced one another and drove the engagement loop of the site.

Just as one example, let’s take from this roadmap that Mark lays out. Events, let’s start there. Events and parties planned on Facebook get tagged with who’s going to attend the events. All of those tags get published out as activity to newsfeed. That drives interest in FOMO among the friend network, so more friends come to the actual event.

At the event, photos get taken. And now there are more people there. While those photos get posted back to the site afterwards and they all get tagged with the people who were there, who are in the photos, all of that becomes activity that then gets published to newsfeed.

That generates more desire from more people who see that on Facebook in the newsfeed to either engage with the photos and like and comment on them, or if they’re hearing about it from their friends and they’re not yet on Facebook to now register for Facebook and get involved in this loop. All of that now gets published back to newsfeed. Now it’s time for the next party. You can see how this just grows and builds on itself over and over and over again.

Ben: It’s funny. As you’re saying all this, it sounds very old hat, almost boring. Of course it does all of tha. At the time it was so revolutionary.

David: Everything I just described was there within the next 18 months.

Ben: Photos I think was late summer of 2005. Interestingly, photos did not originally include photo tagging. Photo tagging was a pretty new concept. I think Flickr may have had it, but the idea that you’re tagging a person and then you can go browse that person’s profile by photos they’ve been tagged in, that wasn’t brand new innovation. That is a mechanic that was not thought of in social networks before. If you think about what Friendster and MySpace were, you could upload a limited set of photos, which by the way you’d have to delete one to add another.

David: MySpace only allowed you to have eight photos maximum at the time. This is—Ben, what you’re describing of person tagging in photos—incredible innovation that drives the whole viral loop.

Ben: And this is a through line through the whole episode. What social media or social networks are definitionally changes every year? Often, Facebook was the one in the early days to push the envelope and say this is what it means to be a social network.

In the later years it was Facebook’s competitors that then they had to adopt that functionality. But the idea that a social network includes an infinite number of photos that you can tag and give XY coordinates on the photo to map to a specific person that is another entity in the social network, that was actually a new component to what it meant to be a social network. As you were saying with newsfeed, that wouldn’t happen for another, what was that, late summer of 2006 or something?

David: September, 2006.

Ben: Yeah, so another year, year-and-a-half after this. That completely turned what social media, social networking was on its head again.

David: Oh, we’re going to talk all about it in a sec.

Ben: To this point, a social network is a set of static profile pages that you can navigate to. If you go to home like the root, facebook.com/ instead of /profile.php, it’s pretty useless. The homepage is actually an uninteresting place to hang out. There is no, hey, let me see what’s going on in my network.

David: Yes. Until newsfeed.

Ben: Okay. But before we get to that, 2005.

David: To your broader point here though, yeah. For the first six or seven years of its life, Facebook was a change maker, defining what social networking was. And then after that it became a change taker. We’ll talk about this when it happens, but again, the mental flexibility of Mark and the company to be like, okay, I’m not going to have pride about that anymore. I’m going to adapt and still win, is incredible.

Ben: Very Microsofty.

David: Very Microsofty. Okay. Even though it takes two years to roll out the full roadmap, photos, events, but especially photos, the 2005–2006 school year is just off the charts for growth, sharing, engagement among college students. Ben, you found an incredible stat about Facebook’s page views, right?

Ben: Yeah, this stat is crazy. By November of 2005, they were getting 230 million page views daily, which means that they had passed Google in page views. Google, the company started six years earlier. The reason is because when you’re on Google, you do one or two searches, and then you go to your destination. On Facebook, you get caught in an hours-long trance of looking at everybody you know and what they’re doing.

David: Which was already the case when it was basically text only. But adding photos, oh my God, every photo is a page view. With such a small user base at that point in time, I think probably certainly less than 10 million users, to have 200 million-plus page views every single day and be passing Google in traffic is wild.

Ben: Let’s say it’s 10 million. It’s 23 page views per day per person. That’s assuming that every single person who is a monthly user is accessing it every day and loading a page 23 times.

In our mobile age now that doesn’t really sound much. I bet the number of photos that someone scrolls through on Instagram is way higher than that. But for the time, that interactivity was just nuts.

David: Certainly Flickr, MySpace, Friendster couldn’t hold a candle to that type of engagement.

Ben: Here’s their user growth. In June of 2005, as you mentioned, they were at 3 million by September of 2005, they were at 5 million users. That was 10X their user base just a year ago in September of 2004, and almost a third of all US college students were included in that 5 million. By September of 2005, they had a third of us college students. Of their 5 million users that they had about 18 months in, 70% were daily active, 85% were weekly active, 93% were monthly active.

David: Wow. Those are just insane numbers.

Ben: I could go on and on and on. But one interesting thing to also point out at this period of time, October of 2005, they were up to 8.3 million users. They were the 10th most visited site on the Internet. But the important part is they were doing $1 million a month in revenue. They had actually started figuring out the advertising business model. Here we are 19 months after founding, they’re no longer burning capital.

David: Well when you say figured out the advertising business model, they just had such a high volume of page views that they didn’t have to figure anything out. You just plug in some crappy ad networks and print money.

Ben: Fair. Yeah, you’re right. Actually I completely misspoke. I would say that they were nowhere near figuring out the advertising business model, but they had a traffic machine.

David: They were able to, through alchemy, turn page views into revenue. But they had not by any means figured out their business model.

Ben: And in this world where startups glorify raising capital, burning huge amounts of money, delaying monetization, and then having these amazing scream and returns when they finally do turn on the money faucet, Facebook was just the opposite. They lost money and not that much money for a year-and-a-half. Then from that point on, they were just profitable.

David: The 2005–2006 academic school year, incredible for the company, everything we just talked about. But there were a couple of yellow flags, shall we say, that popped up.

During that year, Mark of course was already starting to think about, well how can this keep getting bigger? People graduate from college, this is great in colleges, but I don’t want to just build a college site. I want this to be a lot bigger. Obviously, that was what Accel was investing into at that valuation.

They launched two tests throughout the year. One was opening up Facebook to high school students, and the other one was opening Facebook up to workplace groups. Both of them flopped.

Ben: And for different reasons. The workplace ones, they were like, the workplace thing is going to be awesome because it’s authenticated email addresses the same way that the colleges had authenticated email addresses. Most high schools don’t have authenticated email addresses, so we expect these workplace networks to work better.

David: But the problem was going into workplace networks, people didn’t want to share with their colleagues.

Ben: I don’t want some low res party photos of me from last night showing up on my workplace.

David: And I think also just the density wasn’t there. It was violating the principles that got them there of like, this thing needs to be alive and every way we fractally spread out, needs to be bringing that nuclear reactor of what’s happening along with it. Going straight into workplaces, you’ve got 60 year olds and they’re like, there’s going to be nothing compelling for them right now.

Ben: The extreme bear case on Facebook at this moment in time is you started with the group of people who are the most social and the most open to share in their entire lives. You’ve already saturated a third of them. Every single cohort that you add from here is probably going to be worse.

David: And that was a very rational argument. Now, high schools, as you say, failed for a different reason. There wasn’t the same standardized email architecture in high schools across the country. It was just way more fragmented. You couldn’t elegantly set up these private networks in the same way.

Ben: In fact, what they ended up doing, because I was in high school at this time, it was technically a different Facebook. In fact, I will quote the homepage. When you went to sign up, if you went to facebook.com in 2006, the title was Facebook is an online directory that connects people through social networks at schools. Now there are two Facebooks, one for people in colleges and one for people in high school. The site is open to a lot of schools but not everywhere yet. We’re working on it. So if you signed up as a high schooler…

David: You got the crappy one.

Ben: You had to log into hs.facebook.com and it looked mostly the same. It had slightly different features. But there was this weird thing where somebody who was already in college had to invite you. It’s not like anyone could sign up for high school Facebook. You needed a college student who went to high school with you to basically vouch for the fact that you are a real high school student and thus eligible to sign up for hs.facebook.com.

David: Oh man, I vaguely remember this of my younger friends who are still in high school asking for this.

Ben: And the subdomaining thing was that lasted for a long time because the college Facebooks were subdomain as a part of their tech infrastructure. harvard.facebook.com only went to the set of servers that they had set up for that. It made it all really easy from an infrastructure perspective. I was on hs.facebook.com for a long time as a college student.

David: Yeah, because there had to just be one hs.facebook.com for all the high schools.

Ben: Exactly, and it meant that they needed to manage scale and load balance really differently because very quickly the largest network or certainly the largest subdomain was the high school one. So I think within a month or two of allowing high school signups, high school as a whole was much, much, much, much larger, and they had to solve for the technology constraint of what do we do with this? It doesn’t behave like any of our college networks.

David: The academic school year in the core college user segment went great, but once you hit summer 2006, all those kids go on break and college growth stops for the moment.

Summer 2006 was a scary moment for the company. The initiatives to expand outside colleges were not going super well. College growth had stopped for the moment. Yeah, Ben, like you said, very rational argument that, okay, this thing is going to be a really interesting niche site for college students and how valuable is that?

Well, Viacom, which owned MTV, thought that that was worth $750 million, which is what they offered to buy the company for in the summer of 2006. Well, there was another party that really, really wanted to own Facebook at this point in time. That was Yahoo. Yahoo was willing to top Viacom and pay $1 billion to buy the company. Here we are, $1 billion offer on the table.

Ben: And growth is slowing.

David: Exactly. There’s this existential question mark about the company. Growth has slowed a lot. You’ve got $1 billion on the table which is a lot of money, especially in those days. It was shocking six years later when Facebook bought Instagram for $1 billion. This would’ve made the careers of everybody involved.

Mark, of course, controls the company, controls the board thanks to Sean Parker. He doesn’t have to sell, but the whole management team is like, yeah, we should probably sell. I think the board, it’s hard to know. They obviously know they can’t force him to sell. But I suspect if you had asked all of their opinions, and certainly Mark did, they probably would’ve been like, it’s been a great run. We should probably hit this bid.

Mark actually agrees to the Yahoo deal. $1 billion dollars. It’s going to be all in stock. Yahoo’s going to buy the company. In the interim between when they shake hands on the deal and while the docs are being drafted, Yahoo announces their quarterly earnings. They have a bad quarter and the stock drops 20%.

Terry Semel, who was the CEO of Yahoo at the time, said okay, the deal is, for the same amount of stock. Commensurately, the deal is now worth $800 million, not $1 billion. That was history turning on a knife point. That was all Mark needed to say, you know what? Thanks, but no thanks.

Ben: And this part of the story never gets told. The fact that actually at first Mark did not turn down the billion-dollar offer from Yahoo. Mark actually accepted, or at least said let me turn over another card and get one inch closer to the negotiation being final. It wasn’t an outright rejection at first. It was, yeah, come back with the papers.

David: Totally. That is not the way it gets talked about today. But I think actually makes the story even more powerful. It almost really did slip away.

Ben: Well, it’s certainly much more realistic. It’s a simple and powerful story to say they just outright rejected it. But that’s just never how these things go. It’s show me you’re real.

Ben: Okay, analyzing, let’s say he just did turn it down outright. Well, here’s a reasonable way to look at it. If what you want to do is run a great company for the rest of your life, it actually was totally rational. Mark has said publicly, this is like 15–18 years ago, I didn’t have any more ideas as good as Facebook. But that’s way underselling it. There aren’t better companies to start.

Yes. It would’ve been nice to have certainty on a billion dollars or whatever percent of the billion mark owned at that point. also a year before MySpace had gotten bought for $580 million.

David: By News Corp.

Ben: It’s like, geez, okay twice as much as MySpace? But rationally, if the optimization function is I want to run the best company I can for a very long time, almost nobody has started a better company since. If you’re looking at the engagement, you’re looking at the potential, the rational thing actually is just keep running this company because I’ll never discover something like it again.

David: I completely agree with you. Mark has said so much in so many words to you and me personally that he’s not optimizing for financial outcome here. He is optimizing for impact.

Ben: And I think he likes running this company. He would want to go start a similar company if he didn’t have this one.

David: Totally. Also though, to his skill as a strategist, even at this very young age, yes Ben, you and I just made a compelling argument about why you could have believed Facebook was going to top out with the college market at this point in time. Mark knew though, and he had planted the seeds that both newsfeed and open registration were going to be coming later that year.

He was definitely making that calculus here too of, well I don’t know what the probability necessarily is that those are going to hit, but Mark had the belief: (a) if they do, they’re going to be really big, and (b) I think the probability is higher than the people around me think it is.

Ben: Yup, that’s totally fair. Even if you weren’t going to bet on the growth, the user engagement was still currently great, and you had a lot of reason to believe that it was going to get even greater. You compare all the metrics with Friendster and MySpace as time went on and got worse. People churned, their daily active to monthly active would just go down over time, the page loads times would take forever. Especially with, well actually with both.

Friendster just was not architected properly. It was a software engineering computer science issue. Myspace was this weird thing that was born out of this combined media conglomerate, and they just never really had excellent tech architectural talent.

As you looked at not the high level metrics of how many users have ever signed up, because MySpace and Friendster, I think we’re still way ahead at this point, if you looked at how’s this going to play out if it keeps compounding and you looked at the deeper metrics, you thought, oh, I’m running the better company by a lot. Once especially, we got to figure out the business, but I am running the better product by a lot.

David: Okay, to figuring out the business coming out of this episode with Yahoo.

Ben: They say no, the whole management team turns.

David: The whole management team basically turns over the next set of months. Mark gets religion on a couple of things.

Ben: And actually we should say not Dustin, but a lot of the people around the table who were expecting a nice cash payout and now aren’t getting one or are not pleased.

David: What are they going to do? The first thing is I think Mark had always operated the company this way, but now he gets real religion of like we need to get more revenue and we need to focus on becoming truly cash flow positive, so that we are never in a position again where we would even consider doing something like this.

In August of 2006, they do the first partnership with Microsoft. It’s funny that the Microsoft-Facebook relationship in some ways a precursor to the Microsoft-OpenAI relationship today and how that gets built.

Here’s what it’s. Microsoft says, we are going to take over selling all of the display ad inventory domestically within the US for Facebook, and we will give you a guaranteed CPM that we can sell. Then we are going to use this to help bootstrap our online services division, which as we talked about in our Microsoft series, becomes super, super important. Not only for their efforts launching Bing and making that into ultimately a successful business, but even more importantly for Azure that comes out of it.

Microsoft knows they need some scale to bootstrap up and get started with both the ad business but also just like the online services division, period.

Ben: This is a ton of inventory. You just think about the amount of page views that are happening here. Suddenly, Microsoft’s problem is not how do we find enough inventory to sell, it’s how do we go find enough advertisers to actually fill all these slots on Facebook?

David: Microsoft had been trying to do this, they tried to do this with MySpace a little bit earlier and they lost that deal to Google. They really need Facebook.

Ben: And on the other hand, this is a great deal for Facebook because Facebook sucks at selling ads and Facebook has no targeting or anything, so they really shouldn’t justify high CPMs at this point.

David: Outsource it to a good third party network. So this is basically all of Facebook’s revenue for the next couple of years here.

Facebook had made $9 million in revenue in 2005 when they were selling themselves. This is Ben, as you were saying, like hey, they’re starting to turn on ads in 2005. It’s starting to work $9 million. That’s amazing for year two as a company.

Ben: Got to cash flow positive.

David:  Got to cash flow positive. In 2006, with this Microsoft partnership that jumps to $48 million in revenue.

Ben: Oh that’s where the money spigot is.

David: Now we’re talking, and then the next year in 2007, again I think almost all of this is Microsoft, that goes to $153 million in revenue. Okay, we are way out of any league where we consider selling ourselves here. We can control our own destiny.

Ben: Isn’t it wild thinking about, you thought it was crazy to turn down a billion dollar offer or many people thought you were crazy, and then just two years later that is only a 6X revenue multiple and you’re tripling year over year.

David: Oh and by the way, in the second iteration of this Microsoft partnership which we will talk about in just a minute, it included an investment from Microsoft at a $15 billion valuation.

Ben: Hold your horses. Let’s finish this great 2006–2007 arc and then we’ll get there.

David: Yes. Point being though, mark made the right decision for all shareholders of Facebook to walk away from the Yahoo deal.

Ben: It’s an all time high in walking away from every deal ever. He’s made the right decision.

David: Within a year, he made the right decision.

Ben: Yes. Oh yeah, that’s true. That this became obviously right fast.

David: Very fast, very fast. How does this happen? Adam D’Angelo finally graduates from Caltech, joins Facebook full-time as CTO.

Ben: I love that Adam’s our second protagonist here.

David: I know.

Ben: I don’t know if you did that intentionally but, yeah.

David: Well he was right there. They were at Exeter together. I don’t think you can separate it out. And I think here part of the reason why Adam keeps coming into the story is he is a really, really great technologist. Even though he’s young same age as Mark, when he now arrives full-time as CTO, this is when Facebook starts to be taking the steps to be building real technology infrastructure.

Before Adam joins full-time, the team was shipping new code to the website at a high velocity relative to everybody else out there. It was weekly-ish, maybe every couple of days, maybe even up to daily-like.

Ben: Well they were a web company. For the first time in history, you actually could. All these companies with big client Microsoft would ship every three years with a service pack once a year. It’s a whole new era where all you have to do is upload some new PHP code to the FTP server and then boom the application behaves differently the next time someone pulls down the page.

It’s the new technology era of interpreted languages running on web servers in a browser now that products are in browsers meant you could architect your company differently and ship differently.

David: Yes, and when Adam becomes officially the CTO, he’s actually we should probably just ship multiple times a day. We should just be shipping all the time. Also we should probably start recruiting a real top tier engineering team. Because remember, it’s still nuclear winter for startups. Great tech talent is available.

They start going and recruiting like really, really top tier engineering hires, and it’s a pretty compelling offer. We’ve got real revenue from this Microsoft partnership. We’re the best funded startup in the valley. People want to buy us for a billion dollars, we turn them down. That’s all the financial reasons to take it.

For a lot of people too, it was such a breath of fresh air. I think Facebook was potentially maybe the first company, first startup to have an open office plan. Just everything about how they ran the company was different.

Ben: It was the prototypical startup. To this day, early Facebook is still what most startup culture is aspiring to be. I think often without knowing it, everything from the posters on the walls with your mantras and your values to almost having an employment brand that you really care about, cultivating to the idea of we’re having just as much fun together socially as we are working together. I think the modern startup culture, especially when you factor in the shipping everyday open office, why come in here, trains companies to become like Facebook was?

David: 100%. There was Google out there which had a lot of this element, but it was a very different thing. It was very academic, very wonky. These guys were hackers and they shipped, and they were all in it together.

Ben: The other thing that was basically true is they were not really interested in recruiting industry veterans, especially on the technical side. Mark says this in early talks that he was just prioritizing raw intelligence over experience.

David: One of those super intelligent, super talented young engineers with a high slope who joined back in follow five was an engineer named Chris Cox who joined from the Stanford AI lab. Chris joins a team of other young smart engineers led by Ruchi Sanghvi and Andrew “Boz” Bosworth working on Mark’s personalized newspaper product idea.

Rucci was one of Facebook’s very first engineers, and along with her husband, Aditya Agarwal—they run South Park Commons now—Boz was two years ahead of Mark at Harvard and had actually been Mark’s TA in the Intro to Artificial Intelligence CS class that he took, and then he later joined the company.

Ben: Which is actually not how he ended up getting a job. But they did happen to cross paths that way before.

David: Boz went to Microsoft I think before joining Facebook.

Ben: Just for a few years.

David: So the three of them start working on newsfeed. We’ve talked about things like photos and the fundamental architecture of the site. Newsfeed required a whole new level of engineering prowess to get this thing to work. You’re not just going to code this up in PHP. Photos also required real engineering. But newsfeed was pushing the state of the art of what was possible in technology and on the Internet.

Ben: The first question you have to ask yourself, if you’re a good product designer, developer, capital allocator, someone in Mark’s shoes saying should we do this? Is do people want this? If we build this, will it be valuable?

The reason they knew it would be valuable is because the company is data-obsessed, and they watch the analytics like a hawk to figure out what are people doing on our site. They noticed this behavior where people were browsing to other people’s profiles just to look at them and see if anything changed. The user was doing the heavy compute lifting rather than having a personalized newspaper of just bouncing around to a bunch of people’s profiles and saying anything new here?

There was an engineer who did something really hacky because they didn’t want to at first put all the engineering resources into building out something like newsfeed, which is well when something changes for some period of time, we’ll just highlight it in yellow. It’s easy for you as you’re bouncing around to different profiles to just see, oh hey this thing changed. And that totally worked. They watched the lift in that and they’re like oh it’s a good feature.

David: People like that.

Ben: That gave them the confidence of, we should find a way to make it more obvious to you when new updates happen, when things change. Technically, David, for people who aren’t in the tech ecosystem, why is it so difficult to build something like a newsfeed?

Because now we’re all trained to believe that a feed is a primitive that is available to you as a developer when you’re building a product, because feeds are everywhere. Feeds are the core feature of most products. When you hit the homepage it’s some feed. That was not true at the time. That was not true in any product. I actually challenge you right now, think back to 2005. What was a feed on the web? What was an infinite scrolling…?

David: They didn’t exist. This was the first one. This was the invention of the feed.

Ben: Maybe you’ve got Reddit and Digg, and you could make an argument that their feeds, their paginated rankings of stories, so you could see like, oh, what are the most important stories. But even that’s a pretty different fundamental thing.

David: I don’t think Reddit had launched yet. Digg maybe exist, but yes, as you said.

Ben: Reddit was in the first YC batch, which was 2005. It was right around this same time. Okay, so how do you make a feed? Well if you are an algorithms developer, the way you would think of it is, okay, well first I need to pick a point in time. Let’s call that maybe the last time someone looked at their feed. I need to store that timestamp, and now I need to go look at every single profile of someone that you are connected to—this is N—and download or cache all of their recent updates since that time period.

I need to store that somewhere. I need a new place to store a copy of all of this information that lives on someone’s profile or at least pointers to that on everybody’s profile. I need that to happen for every user. Now it’s N-squared.

Every single person on the entire Facebook needs to have something running in the background that is looking at every other person on Facebook since a particular time. Then that compute and storage all needs to happen fast enough such that by the time they want to go check the feed again, it’s happened again.

Obviously, now that happens in real time, and I think it was something like every three hours there was like a new batch.

David: It was four times a day when it initially launched. There was a 4X a day refresh.

Ben: Okay, so every six hours.

David: It was a cron job because they didn’t have enough memory to run it in real time.

Ben: That’s right. They needed to happen on separate boxes to run this process, cache the results, and then when you loaded your newsfeed, go fetch them. This is a whole new application using the same data that the company has to build in order to make newsfeed happen.

David: What you just described is all the technical work just making a feed possible.

Ben: On the back end. This doesn’t contemplate any of the front end design or engineering or incredible permutations of how to display this data when it does come back, given a massive combinatorial problem of how might this data come back.

David: Exactly. Then you need to make the feed actually compelling. Like I don’t really care about somebody who I’m tangentially connected to what they had for lunch today, but I really care about a photo of Ben at the Acquired meetup.

Ben: So now you’re telling me you want to rank order it by something other than.

David: Chronology.

Ben: Make another pass and figure out what I think is going to be the most interesting to you. Which on its own is an incredibly difficult computer science problem. What is interesting to you? What data should we use to inform that decision?

David: Totally. Who are your close friends? Who do you care about the most? What general news do you care about the most? What types of stuff do you care about?

Ben: Do we have to put weights on every relationship in this entire complex friend graph between every single entity and how close they are, and then re-rank that very often? Eventually, yeah.

David: The team spends the better part of 2006 working on this. By September, it’s ready. By September 5th, 2006 just in time for the new school year, they launch it. People noticed because they launch it to everybody right away. This is a massive paradigm shift.

Ben: It’s not opt-in or anything.

David: This is Facebook changing the game of social. Not opt-in, everybody gets it right away. They get 30,000 angry emails to support on the first day from users who are really upset about this. Ten percent of the entire global user base signs up for a Facebook group called Students Against Facebook Newsfeed. I’m pretty sure I was part of this group. The irony here is they had only just launched open groups across school networks days or weeks before. They enable the tool of their own vitriol here.

Ben: And also, guess how people are finding out about this group?

David: Through newsfeed, yes.

Ben: This is the great irony the whole situation. There’s a literal panic in the streets. There are people protesting outside the office. There is somebody trying to use a crane to get into the third story. Basically a TV truck is trying to cover what’s happening in the pandemonium.

At the same time, despite everyone telling them, I hate this thing, it’s the worst thing ever. If you look at the analytics, people love it. People cannot get enough of scrolling through the new newsfeed. The reason they’re really upset is actually quite interesting. They’re saying, oh it’s sharing this with my whole network. Okay, but you updated it on your profile, which has always been public to your whole network.

I think Facebook for the first time stepped in it and realized, oh, even though technically this data has always, we didn’t change how public or private it is, it’s just as accessible, people react really strongly when you change the ease of obtaining that information or whether it feels like you are pushing that information out versus someone is pulling by going to your profile and viewing.

David: It went from pull to push. So it’s the start of the new school year. We’ve just come off this tumultuous summer. Growth had slowed, walked away from the Yahoo deal. There’s this revolt in the streets against newsfeed, which to Mark’s mind along with open reg, which was supposed to come two days later of opening up Facebook to anybody, they delayed that because of the newsfeed reaction.

But to Mark’s mind, the two of these things are the big growth levers to reignite growth for Facebook. The board, the management team, everybody’s like, all right, we got to roll news feedback. People hate it. This is PR 101, we apologize, we roll it back. To add a kicker, newsfeed actually significantly hurt Facebook’s revenue because where do you think all the page views were coming from and all the refreshes?

Ben: Loading a new ad.

David: Yup, loading all the profile views and clicking around. Page views are actually going to go down in this new paradigm. Every reason is aligned to roll this back. Richie, Chris, and Boz though are looking at all the data and they’re like, holy crap. Engagement is through the roof.

Ben: People can’t get enough of this.

David: People say they hate it, but yeah, it looks like they can’t get enough of it. Mark decides that he is going to write a post on Facebook about this, and it’s titled Calm Down, Breathe, We Hear You. He announces that they’re going to launch a set of controls for you to control what of your activities get published to newsfeed and what don’t.

Ben: Which, by the way, sounds like something that Mark would say today. I think this 2006 era is the last time and then you would’ve had a 17-year break, and now we’re getting that Mark again. You can’t imagine Mark circa 2018 saying, calm down, breathe, we hear you.

David: It’s a brilliant strategy of like, hey, I acknowledge that this was surprising. We didn’t handle the rollout right. I’m seeing that you all actually like this in the data. Let me give you some controls, so that what you’re really worried about, you have some control over, and then let’s just see what happens.

Within two weeks, it’s like magic. Everybody just gets used to it, and that this is now the way Facebook and social media operates, and engagement continues to skyrocket.

Ben: But it’s funny. I did just pull up. We’ve been getting a lot of feedback about newsfeed. We think they’re great products but we know many of you are not immediate fans and have found them overwhelming and cluttered. I don’t think clutter was what people were complaining about.

David: A little bit of redirection there.

Ben: It’s funny. We didn’t take away any privacy options. Your privacy options remain the same, the privacy rules haven’t changed, none of your information is visible to anyone who couldn’t see it before the changes, blah-blah-blah. It is interesting. It’s true and not relevant.

That’s not what people are mad about. People are pretty aware that this is the same information.

David: And I think to set a theme here that we’re going to come back to several times, the definition of what social media is is actually very fluid, and it changes in the consumer’s mind.

Ben: I’d even say before this, there wasn’t social media, or at least Facebook wasn’t social media. Facebook was a social network, but this was the first time they introduced a media component, a thing you would read.

David: Inspired by a newspaper. I think when these paradigm shifts happen, people get upset because their expectations are being violated. It actually doesn’t really matter what the privacy is or isn’t. It’s the expectations.

Ben: I’m glad you planted the seed because this will come back over and over again in their history of people now feel differently, and the product needs to change with those societal expectations in order for people to not be upset about it.

David: Unfortunately for the moment, I think the lesson that Mark and the company took from this experience was, well we actually know what’s best and the user base will just get used to it whenever we make a change. It just happened to be that that was true with newsfeed, but that wouldn’t always be true.

Originally, the two big initiatives, newsfeed and open registration, were supposed to basically launch together. They chose to do newsfeed first because it was going to be: (a) a little bit easier to roll out, and (b) more valuable to the core college audience that was just starting back for the new semester, the fall semester at school. They launched that first.

The original plan was open registration, Facebook is now open to anybody in the world, was going to come two days later on September 7th. They obviously put that on the shelf for a long time, 2½ weeks. At the end of September, open registration launches, anyone can sign up for Facebook.

This is another major, major change. Before all these networks were siloed, you had to be part of a verified email address network to join Facebook. Now anyone can sign up.

Ben: This is the beginning of uh-oh, my mom is on Facebook.

David: Exactly. Which ultimately would be a problem because that’s what created room for Instagram, for Snapchat, et cetera. But that would actually be down the line.

In a weird perverse way, I think because of all the controversy around newsfeed, when open reg launched, there wasn’t the same controversy. People were already desensitized or they had just been through newsfeed and they’re like, all right, yeah, whatever. Nobody even really noticed that much, especially because the product experience didn’t change for the already engaged users.

Ben: And to the extent that you did feel that privacy had changed. You’re now used to anything I put on Facebook gets broadcasted, so. who cares if more randos come in? It’s already getting broadcasted.

David: Yup. And now with newsfeed, there is this mechanism that makes sure like, hey, even as randos from your perspective, join Facebook, you’re still seeing the updates of people you care about.

Ben: Well I don’t think at first when it was this pure chronology.

David: Yeah, I guess that’s a good point. As they make it more algorithmic.

Ben: But this was still a period in time where your friends were your friends. Facebook was only two years old, and for the vast majority of users they joined in the last year, they didn’t have anybody who wasn’t really their friend as a friend.

Now I just have to treat Facebook posts as if they’re public because the group of friends is aged. But in that period of time, you could trust if something was getting published to your friends that like it’s just going to your friends.

David: The age of the network itself was only two years. Your friendships hadn’t shifted that much. Well, open reg despite high schools and workplaces not working well, open reg works really well.

Ben: Facebook has a brand at this point. It’s the best social network. It’s the fastest-growing social network. It’s the one that all the college kids will always be the cool people in any society at any time. That’s the age group of trend setting. They have conquered that market. High school kids want to be like college students and people not in college want to be like college students.

David: Yup. Over the summer Facebook was adding 5000–10,000 users a day. By late fall after launching open registration, they’re now adding 70000 users a day.

Ben: Wow. By fall of 2006.

David: Growth really picks up. This brings us now to spring of 2007 and the final piece of Mark’s original product roadmap: I’m bored AKA platform. Platform is this forgotten thing right now. It’s Farmville, it’s Zynga, it’s apps on Facebook.

Ben: It’s quizzes.

David: It’s quizzes, yes.

Ben: That’ll come back up later.

David: That’ll come come back. For a period of about four years from 2007 to 2011, this is the most important thing in the company, and this is what everybody believes the core and feature of the company is. Facebook has become a platform. This is the goal of a technology company, become a platform, be like Microsoft. This is now Facebook is like Microsoft. We are a developer platform.

Ben: And there was this incredible tension, at least it seemed like the tech media wanted to play up this tension. Are they an advertising company or a platform company? Because they’re super different business models with very different incentives, and users need to know how to think about it. It was at the same time Sheryl had just joined the company or was about to join the company?

David: The next year in 2008. Yeah, not yet.

Ben: They really hadn’t gotten serious. She led them to figure out, hey, let’s survey all the business models, do some work on each of them and figure out and commit to being an ads company. At this point in time, they were showing ads, but the whole leadership team, at least I’m convinced, believed we are on the way to becoming a platform.

David: 100%. May 24th, 2007, Facebook holds its first developer conference to launch their developer platform, F8, which again, let’s take a step back, this whole idea is wild. A developer platform on the web. The web is a developer platform. Facebook is now becoming a platform on top of an open developer platform.

Ben: But they have two important things that as an application developer you’re interested in: (1) user attention, the same way that Microsoft had user attention with PCs because they had the install base of Windows, and (2) a whole crap ton of data about each person that you could then build into your application to make it really rich and feel personalized.

David: And it turned out that that was really attractive to developers. At F8 they announced that the Facebook user base is now over 20 million and it is growing by over 100,000 users a day, so growth is accelerating.

Ben: Wait, it’s 20 million total and growing by 100,000 a day?

David: Yes. Open reg had just launched a couple of months before. This thing is really taking off.

Ben: Oh my God. Every 10 days they’re adding another 20th of their user base, and they’re growing by 5 per, what is that? It’s like 7% every 2 weeks. It’s like 3.5% a week?

David: Yeah. Pretty good. At a scale of over 20 million already. It is the sixth highest trafficked site in the world, and they’re announcing to all the developers in attendance, the Facebook API, Facebook Graph is now open to you. You can build apps and publish and run them natively on Facebook.

For users, and part of Mark’s original product roadmap from 2005, this is huge. Until this point in time, Facebook was about digitizing everything that happened in your offline life—your real friends, your real parties you went to, the real photos you took. There wasn’t anything detached from your offline reality to do on Facebook. Now all of a sudden you can play games on Facebook, you can take quizzes on Facebook, you can use apps on Facebook, and you can do them all with your friends.

Ben: It’s pretty compelling. I remember playing, was it Scrabulous?

David: Yeah.

Ben: I think I was playing with my sister, with my grandma. This is a great way to do things digitally with the people that matter to you.

David: I think it’s also an early foreshadowing of the job that mobile and the smartphone would really do in people’s lives, which is hey, there’s actually a lot of white space where you’re just bored during your day. Mark’s original title of the feature of I’m Bored was Perfect. This is the cure for boredom. It just so happened that it was on a desktop, so when you were bored at your desk or bored at your home. Mobile then opened that up to hey, I’m bored anywhere.

Ben: And from a business perspective, it gave Facebook another stakeholder. Hey, developers are someone who can keep people on the platform longer, so we can show them more ads. They might do their own advertising, so they might drive traffic to Facebook that they could help grow the core platform itself. It gives us more lock-in as people develop for us and as users seek out applications on us. They’re hoping for that Microsoft playbook of platforms get really epic lock-in.

David: Oh, and by the way, developers on the platform will probably have their own business models where they’ll make revenue, especially if we ultimately introduce our own virtual currency, we can probably make revenue from their revenue too.

The internal goal of the company was to get 5000 developers in the first year who would be making social apps on the platform.

Ben: How many did they get?

David: They got 5000 in two days. developers went freaking nuts.

Ben: It was a great keynote.

David: It was a great keynote, yes. They went nuts because yeah, Ben, as you say, oh, distribution to a lot of people, and then bringing along your friends and data on them. That’s a very powerful incentive as a developer to then go make an app or a game for the platform.

Ben: You’re building a web app that has an unbelievably rich set of data that you can hook into. The thing that I was, I don’t think I could have put into words then but I can now, is has a successful scale durable platform ever been built that wasn’t an operating system?

It always felt weird to me in that moment that Facebook thought they could be a platform, because I was like, well it’s a website. It’s a web app and I have a profile, and I have all this rich information, but it’s not running on my device.

David: Well Ben, turns out you would’ve been right.

Ben: If only I had the words. But what do you do with that information? What was I? Going to be bearish on the company? That would’ve been a super wrong call.

David: Well this is the amazing story that we’re telling here of they keep surviving and thriving despite not having the operating system.

Ben: Listeners write in, if you can think of one, a successful, scaled, durable platform, where the platform is able to make money and they’re able to make a lot of money for developers, for people on the platform that is not an operating system that runs on…

David: Hardware, either captively controlled hardware like Apple or Open hardware like Android and Microsoft.

Ben: And this gets into the question of what is a platform, because let’s say I make a web app that I run on my own website, I do my own marketing, I do my own monetization, but I let users authenticate Facebook so I can pull some information out of their profile. It’s like it’s not really built on the Facebook platform. It’s not like the core APIs that enable my application to run are using Facebook’s APIs. I’m using Facebook’s APIs to grab some data. The core API set that allows it to run is the browser.

It’s almost like if you think about the intermediation layers, Facebook was trying to build a platform on top of a browser that was a platform on top of an operating system that was a platform. The reason why we aren’t all using rich Facebook apps all the time today and think of it as the default platform, it was just too many layers of abstraction away from the hardware to win.

David: I think that is totally right. But for the moment, thousands and thousands of both indie devs and venture backed companies flock to the platform.

Ben: But what this does tease out is Facebook, and now we can talk about this 15+ years later, has a weak position for launching a platform since they don’t control the OS or the hardware. They have to make a little bit more of an appealing sale to a developer. That includes big distribution for you, but it also includes a ton of access to data.

I remember being an early Facebook developer. After a user authenticates, looking at that JSON and being like, okay, so it’s their whole profile. Oh it’s their whole friends list. Wow, I get a lot of stuff here. Facebook was incentivized to do that because they almost had to sell harder than other platforms who controlled hardware historically would’ve had to.

David: It’s a super great point. Now also, the reality was at the time I think developers were getting plenty of value just out of the distribution, and that was coming from newsfeed. If a user starts using your Facebook app and then the activity that they’re doing on your app is getting published to the news feeds of all their friends…

Ben: Oh yeah. You could build a whole Zynga on that interaction paradigm.

David: Oh yes you can. There’s this great quote from Josh Elman in the Steven Levy book. Josh says, “If you’re a developer and you can get someone to bother 10 friends to get 1 more user to join, you’re very happy because you just got 1 more user.” Facebook, though, has nine other people who’ve just gotten bothered, and he doesn’t say it, but are quite unhappy about that.

Ben: And that’s the trade-off they have to weigh. That’s their in the type of platform they are trying to build. They have set up that incentive set and they need to figure out what to do with it.

David: Yup. So for a year or two, that is the state of play, Facebook, brilliantly for them, eventually pulls back on organic newsfeed distribution for apps, says hey, we got to stop this. We got to take care of the problem.

Ben: You don’t need to know every time there’s a new mafia war move by so and so showing up in your newsfeed.

David: And that does effectively kill a large percentage of developers and apps on the platform.

Ben: But it’s the right long-term move. That was the right thing to do for Facebook with a 20-year view.

David: Absolutely. And it doesn’t kill the platform though. It actually makes it even more valuable to Facebook because for the developers and apps who have gotten to scale, they look at this and they say, my organic distribution just got kneecapped. But Facebook is still this pretty powerful platform. What if I just buy ads? How well would that work? Can I generate positive ROI if I buy ads to effectively accomplish the same thing natively on the platform and drive usage back to my app?

Turns out that there was a very profitable arbitrage there still for many years to come. And for Facebook they’re like, well hell yeah. Okay. We’re monetizing the platform through user engagement and ads that we’re showing on the right rail while people are using Facebook apps. We are also monetizing through currency and commerce that is happening in the apps where we’re taking a cut of that through payments.

Now we’re going to monetize platform in a third way, which is the biggest scale, developers and app distributors on the platform are going to use our native advertising as their primary form of distribution. You can see why for a couple of years here Facebook is like, well this is it. We’ve won. We are a platform. This is it. Then mobile comes along.

Ben: Oh mobile. That is quite the chapter in the Facebook story. But before we do that, now is a great time to tell you about one of our favorite companies, the climate-aligned AI infrastructure company, Crusoe.

David: Yes. Crusoe is a vertically integrated cloud platform built specifically for AI workloads that was recently named the Gold Standard of AI cloud providers by Dylan Patel over at SemiAnalysis. Something that’s new and really cool since we started working with Crusoe last year, they’ve opened up the raw metal in their data centers to customers.

Let’s say you’re a large enterprise who wants to run your own infrastructure and not use a cloud layer. You can actually now do that with Crusoe directly as a data center customer, which several of the largest companies in the world are now doing.

Ben: They’ve totally reimagined traditional data center architecture to support the huge power cooling and compute density needs of AI.

David: That’s important because power demand in GPUs is increasing dramatically, which means the traditional data center design and engineering of the hyperscalers is no longer optimal. Crusoe infrastructure built from the ground up for GPUs with elements like high density racks, direct liquid to chip cooling, that enables them to support the most demanding AI workloads that traditional data centers just can’t handle.

Ben: And at the same time as GPU cluster sizes continue to increase, there’s an ever increasing demand for energy. Crusoe has 15 gigawatts in its development pipeline, which is an astronomical amount of power. Their Abilene Texas facility alone has over 1.2 gigawatts planned, which will make it one of the largest clusters in the world.

David: And it’s also not just about the amount of capacity in the pipeline. It’s about how fast it can come online. Crusoe’s team has decades of experience constructing and operating data centers which enables them to bring these new data centers online super fast.

Ben: As many of you already know, Crusoe sources the energy for these data centers in the most efficient and climate-aligned way in the entire industry—through clean, low cost, and abundant energy that otherwise goes to waste.

For example, in oil fields where natural gas is flared in congested parts of the grid where renewable power is curtailed or other areas where energy is stranded, which actually accomplishing that is a crazy hard thing. Crusoe’s energy-first approach means they can build data centers in some of the most challenging locations on earth, bringing computing to the energy rather than the other way around.

David: The net of all of this is that Crusoe can provide nuclear levels of power for far less cost than other providers and with low or in some cases actually negative emissions. That’s super important because the biggest bottleneck to AI progress is actually energy. It’s not compute.

Ben: Anyway, they’re just a great company. We’re super proud to work with them and to be investors. To learn more about Crusoe, you can go to crusoe.ai/acquired or click the link in the show notes and just tell them that Ben and David sent you.

David: Thanks Crusoe.

Ben: All right. David, we’re here in the era of mobile. This is January, 2007. Steve Jobs just announced the iPhone. The whole world changed, right?

David: Well actually ironically, no. Or at least not for quite a while. Yes. January, 2007, Steve Jobs announces the iPhone. People forget though, it took five years for mobile and smartphone adoption to really ramp up and actually change the world. There was no SDK in that first iPhone and iOS. That didn’t come till 2008.

Ben: It was $700. Android was a Blackberry clone at the time, not an iPhone clone. The DROID DOES campaign wouldn’t be another year or two.

David: I think two years. I think that was 2009 if I remember right. Hell, Facebook platform didn’t even launch until May, 2007, so after the iPhone announcement.

For the moment, platform is rocking and rolling, like we were just talking about. Facebook is becoming the next Microsoft, which brings us to the second Microsoft partnership in October of 2007. Just five months after F8 and platform launch, Facebook now has all of the leverage. Dan Rose has some great stories on Twitter about how this negotiation goes down.

Ben: You spent time with Dan, didn’t you? Prepping for this?

David: Yup, I chatted with Dan about this. This is great. Microsoft perhaps obvious now in retrospect to everybody, really just wants to buy the company, but Mark’s of course not going to do that after the whole Yahoo experience, et cetera.

Microsoft is also happy though to just keep the partnership going and expand it, because: (1) it is really helping them spin up their online services division and get really good at online ad sales and ad serving, and (2) they know they made a huge error in missing search and letting Google get big. Social sure seems like it’s going to be the next search-sized platform, so they’re happy to just be in bed with Facebook in a way that they never will be with Google.

Ben: As long as they can keep it going. For them, I think the thing that shocked them about Google is, oh my god, the business model of the web is advertising. Search and browsers and everything are monetized by advertising, and that’s completely orthogonal to our traditional license-based business model.

Social sure looks like where all the page views are going and is going to monetize exactly the same way as search. Actually, the thing that’s important to them is locking in those page views, to the extent that they can participate in this market. It has to be either we own you or we are the long-term ad provider, which they didn’t end up being, or we’re a big equity holder.

David: Enter this second partnership negotiation. October, 2007, Microsoft comes down to Palo Alto. Facebook tells them that hey, Google is actually really interested in taking this partnership over. Larry and Sergei have been banging down the door, and we’ve been trying to hold them off, but they’re coming to meet with us tomorrow to talk about switching our Facebook ad serving partnership over to Google.

This has been great. We love working with you all. We want you as our preference, our preferred partner. We want to keep it going with you, but we need to get the deal done tonight. They lock themselves all in the conference room, they start negotiating, it’s getting late, 10:00 o’clock goes by, 11:00 o’clock. The Microsoft team, by Facebook standards, is all old guys. They’re getting tired.

Ben: God.

David: Old guys, like 40, like oh it’s so old. Then at midnight—this is amazing, you can’t make this up, Facebook probably planned this—all of a sudden, this really loud house music starts blaring with heavy bass in the office. The office is hopping. Everybody’s there. The whole company’s there. It’s midnight. The Microsoft guys are like, what’s going on? And the Facebook decides, oh yeah, we have a hackathon scheduled tonight. Actually, Javier Olivan had just started at Facebook.

Ben: Who’s now COO.

David: Who is now the COO of Meta, had just started at Facebook as an engineer. He had organized this hackathon to begin work on internationalizing the site. Tonight is the beginning of internationalization, which is a critical, critical part of Facebook’s next chapter of growth. And it is beginning at the same time that this Microsoft partnership is happening in a conference room.

Ben: And didn’t this Microsoft deal end up being about international?

David: Yes, it did. They keep going. At one point during the night, one of the Microsoft guys, according to Dan, looks at him and says, wow, this is awesome. This is just like the old days of Microsoft.

They finished the negotiation at 6:00 AM. They announced the new partnership that very same day as promised with the threat of Google. Microsoft becomes the worldwide exclusive third-party advertising network for all display ads on Facebook. It was domestic in the US before, now it’s international too and domestic everywhere around the world.

Facebook, of course, can still also sell their own inventory themselves, but anything that they don’t want to sell themselves or that is excess inventory, Microsoft gets exclusive access to be the only ad network where you can access it.

The big kicker, Microsoft is going to invest $240 million in Facebook at a $15 billion valuation. This is 13 or 14 months after the failed Yahoo deal for $1 billion.

Ben: That is the thing that no one talks about here. I’m so glad you brought this up because before they started this negotiation, the attempt was to buy Facebook. The highest offer floated. We talked about this in the Microsoft episode. It was a complicated set of deal terms that basically netted out to a $24 billion offer.

That was less than two years after the $1 billion offer that everyone talks about with Yahoo. Oh remember the time where Mark Zuckerberg turned down $1 billion? Less than two years later he turned down $24 billion. They were really good for the money. It’s Microsoft.

David: But turned it around into an investment at $15 billion.

Ben: Yeah. It’s crazy.

David: So the great thing is like all great partnerships, everyone makes out wonderfully here. Except maybe the venture capitalists in Silicon Valley because this just breaks everyone’s brain.

Ben: It was the high watermark for Facebook’s valuation for a while because then the great financial crisis would happen after this, valuations would all reset, and then Facebook’s next deal would get done (I think) with Yuri Milner at $10 billion.

David: $10 billion valuation with Yuri, yup. And Yuri would actually also buy common stock through a secondary tender offer in that. His dollar cost average would get down to more like six, so he got a screaming deal.

But for the moment, there are no comps to this, ever. Dot-com bubble, you name it, never has been a comp like this for an investment in a private technology startup. The closest comp that I can even think of was Google’s market cap at IPO was $23 billion, so higher but not that much higher. Google at IPO was a $2 billion revenue run rate spitting off hundreds of millions of dollars in free cash flow annually.

Facebook at this point in time is basically break even and doing $150 million in revenue, most of which is Microsoft itself. All that said, Microsoft is a home freaking run. We’ve been saying all along. Helps them spin up Bing doing all the ad serving for this, getting into real online services.

And then Microsoft is not a hedge fund as we talked about on that episode, but that $240 million investment, by the time they start selling it down years after Facebook is public, ends up being worth $8 billion.

Ben: Wow.

David: And then maybe the most important benefit to Microsoft out of all of this is Microsoft and Facebook have always had a good relationship, unlike Microsoft has with Google and Apple.

Ben: I remember Facebook events, like when you look at the page, it was a Bing map, not a Google map. That always felt emblematic of the partnership to me. Every single place, it could be a Microsoft product. It was.

David: Yes, it wasn’t a Google product.

Ben: Yup. Okay, interestingly here we are in 2007, Facebook is still saying we don’t need to be in control of the core revenue creator for us. We think platform is the future. Microsoft, you’re our preferred partner to handle making the money.

David: I think there are two things here. Yes, I think that is true specifically at this moment in time. But this is where again, Mark is just a master strategist, calling it all the way back to Civilization and the 4X strategy game. He’s got multiple bets he’s placing on the chess board.

One, as you said, is platform. Platform is rocking and rolling. We think we’re building the next great technology company and the next great business model. Just on platform. Also, yeah, we’re a social media company that involves a lot of page views and engagement. The right way to monetize that is advertising. So yes, we’ve just continued this big Microsoft partnership, but we have the right to sell our own inventory and we should probably start building that muscle too.

Ben: The other thing that’s happening here is Mark still thought ads were gross at this point in history. The things he cared about were product and engineering and design, and he wanted to outsource everything else or at least have someone else at the company think about it.

Having Microsoft take care of the ads was (I think) in his mind a win-win. That way the commodity stuff can happen elsewhere and I can focus on the thing that really matters. The company really did not have the insight yet. Oh, we can do ads different and better than anyone has ever done them before.

David: Well, I’m smiling as you’re saying that because in practice for this moment, yeah. In reality they were trying to build their own native Facebook advertising unit, so to speak. It was just Beacon.

Ben: Yeah. Beacon is one of the most epic failures in Facebook’s whole history.

David: Okay, Beacon. Well first off, you were totally right and I think you were foreshadowing Sheryl Sandberg there, of Mark wants somebody within the company to just run, manage this thing, and build it great and world class. Before Sheryl, do you know which Facebook executive led Beacon?

Ben: I do not.

David: Chamath Palihapitiya.

Ben: No way. He always talks about the growth team. I never knew he was in charge of Beacon first.

David: This was Chamath’s big thing before growth. Actually, the epic failure here of Beacon leads directly to growth as we shall see in a sec.

Now, what was Beacon? Like you said, Mark is allergic to traditional advertising, but he’s like, well we have this incredible social activity happening on Facebook and we know through newsfeed that people engage with and they love it. What if there were a way for a native ad baked into the very fabric of the product itself that brands could control and monetize?

Well, if we gave brands a way to essentially boost how people are already engaging with them, that might be the way to do this. What this ends up being in practice is brands publishing your friend’s e-commerce activity into your newsfeed.

Now, to be fair to Mark and Chamath and the company, we didn’t know yet what the right advertising unit on Facebook was going to be. And this is probably as good an idea as any because like the core thing that you do on Facebook is you engage with your friends and you engage with their stories on newsfeed. There’s no evidence really that people would want to do anything else but that, so you’re trying to think of how we shoehorn advertising and brands into this?

Ben: I actually don’t buy it at all. The obvious thing is show an image that people can click on and take them to a website. It’s display ads. I don’t understand why they needed to try to way overthink it and say our ad format has to be something that no one’s ever thought of before. Just offer advertisers to do the thing that they know how to buy.

David: Well, I was trying to think of how to be the most charitable possible to the company and how they can come up with it. But yeah, I have next in my notes here, to be clear, this was a truly, truly terrible head up your rear end idea.

Ben: These are actually two different things, I think. There was Beacon, which was JavaScript that an advertiser could embed on their website on an e-commerce provider’s website that would do exactly what you’re talking about, publish into the newsfeed purchases that people were making.

This was at the point where you wanted to publish all sorts of interesting different things on social media. I don’t think we knew for sure that purchases weren't going to be one of them. People still keep their Venmo history public. You don’t know what people are going to do.

The second thing that I think is in your description a little bit is the idea of social ads that brands could take interactions that happen on pages and boost them, but they couldn’t just take out a regular old ad. It was this weird thing where you could only advertise as a brand to people where someone in their network had interacted with your brand page.

Pretty convoluted. I get that it’s this super natively social thing, but again, it did feel like they started in this place where they’re like, let’s get way too clever for our own good first and then work backward to the most basic ad unit.

David: Totally agree. That said, as soon as it’s live it becomes clear that: (1) nobody wants either of these things..

Ben: Users don’t want it. Users are confused. Certainly there’s instant blowback against their publishing My Purchases. Someone (I think) had a engagement spoiled by a diamond ring getting broadcast.

David: Oh yeah. I was going to say, this is (2) the privacy implications here actually are horrific, like spoiling engagements. That’s the wholesome, horrific privacy thing. You can imagine the non wholesome publishing stuff that you’re buying versions of this.

Ben: And then of course advertisers are confused and people don’t really understand what’s an ad. What’s not an ad. They just burned credibility everywhere by launching both of these things.

David: Totally. Mark and the company is coming off the newsfeed experience where they’re like, yeah, if we just give it enough time, users will get used to it. They let it run for a couple of weeks. They run the newsfeed playbook of like, yeah, yeah, yeah, we hear you. but you’ll get used to it. Here’s some more controls.

Ben: Oh, the social ads was years and years. But the purchases, you’re right. That was just I think a couple of weeks.

David: Yeah, for Beacon. Ultimately after a few weeks, they do acquiesce and they say like, okay, we will turn Beacon off. We’ll make it opt-in as opposed to opt-out. You can have a way to completely turn off all Beacon tracking permanently from your profile.

A lot of people do that. Beacon activity basically drops to zero and then two years later it’s completely killed off. This sets the stage for Sheryl joining the company. I think Mark’s takeaway from this is like, okay, maybe I don’t understand advertising as well as I thought I did. I thought I didn’t like it, but I thought I could be clever and engineer this new thing. Actually, let’s bring in somebody who really, really does understand this.

Sheryl was the perfect person coming from Google. Sheryl joined Google in 2001 right as they were figuring out AdWords.

Ben: That’s three years after the company was founded.

David: Totally. Well and the company didn’t figure out AdWords for a couple of years. That was the key moment. Then Sheryl built and ran the whole self-serve business at Google, which was the most important and the most technology-enabled part of what they were doing.

Ben: She had built the greatest digital advertising system in the world. Mark was like, oh, I’m trying to build the greatest digital advertising system in the world. Who could I ever get to do that with me? And wouldn’t it be great if that person was also a good manager and leader who could teach me how to manage and lead?

David: I think that really is the “deal” they make. They say, and Sheryl agrees too, it’s not like, hey, Mark said you’re going to take this side of the company and I’m going to take that side of the company. She always reported to him. The whole company was his company.

But Mark is really focused on product, engineering, and importantly platform. Sheryl takes ads, Sheryl takes operations, et cetera. When she comes on board, like you referenced this earlier, Ben, there’s basically no ad targeting that’s happening natively on Facebook. There probably is targeting that’s happening through the third party Microsoft advertising network, but none of that is like technology that Facebook’s building or value that’s accruing to Facebook.

The first thing Mark and Sheryl did when she joined was, okay, Beacon was such a disaster. We’ve got this great partnership with Microsoft. Let’s just do a full exploration. Should we be in the ads business at all?

Ben: She definitely famously led the exploration of what business are we actually in here.

David: Ultimately they decide, yes, we are in a media business, thus we should be in the advertising business. We should do it right and we should make it work. Ultimately that is going to be building a really, really, really great targeting engine, which is the thing that Facebook is uniquely capable of doing.

So though, back to here where we are in 2008, two seemingly unrelated but about to get very related things are going on: (1) After the Beacon fiasco when Sheryl joined, Chamath needs a new job at the company. Total credit to Mark, to Sheryl, to Facebook as a company. I think at most places like Chamath would’ve been fired immediately for what happened with Beacon. But the culture at Facebook is like, nope, we move fast, we break things and we’re always on offense. This didn’t work and let’s find the right thing for you to do. Thank goodness, because that brings us to (2) right around this time. Once again, user growth slows at Facebook.

Ben: It’s funny that I don’t feel like in the public eye we really knew these things, But as you dig into the company history, you’re like, oh, there are these moments in time where growth really did stall out. Chamath has said on stage since then, oh, I totally thought that growth was done. There are a few things to note about the growth team. One is, it was effectively the first growth team.

David: I think it was the first.

Ben: Everyone who wants to get into growth now, I don’t think really fully grasps that is a brand new discipline invented in 2008. And the way they defined it at Facebook was there’s marketing, there’s product, but marketing and product both touch customers in different ways, so the left hand needs to talk to the right hand.

For example, you should not have a marketing team that is sending out emails through an email marketing system, and a product team that is sending out notifications to users through the product with no notion of how they’re speaking to each other. You should: (a) unify those efforts through one team, and (b) that team should live within product or at least that team should be tightly coupled to product, with the general belief or reason for being that your product is the best lever to grow.

No amount of marketing you could ever throw at something that is not integrated into your product will be as effective as your product doing a good job with features, hitting the right users with the right message and the right value at the right time in a native way to the core feature set of the product.

You’ve got Chamath, you’ve got Alex Schultz, you’ve got Naomi Gleit, and Javier Olivan. you’ve got this core early team, it’s four people, and then it expands to six or seven folks, that are formed really on this agreement between Chamath and Mark of, we are going to have this dedicated growth team and our mission is going to be to grow, Facebook using Facebook itself, not through extrinsic measures.

David: This was an incredible insight and was so perfect for Facebook because I think this growth insight of growth through product can apply to any company, but it especially applies to a virally growing social network like Facebook.

Ben: And there are these interesting things that make it tick. The first of which is this has to be the most data-obsessed team in the company. Every team should be data obsessed, blah-blah-blah. The growth team is really the one who discovers, oh here are the obvious places where users are deriving value. Here are the obvious places where users are getting confused. Here are the obvious opportunities to find new users. Here are the obvious opportunities to reduce friction. Analytics is the answer for that.

I think that there are a lot of really interesting stories, especially around internationalization, of the growth team and partnering with other product teams around the company to say, what is an engineering and product approach to something that traditionally has been done other ways? Like the way most people would translate their product is by hiring translation servicing firms or by hiring a dedicated person to go through string by string and edit.

Well we’re Facebook, we have all these users. What if we just, when we’re launching in Spain, surface different words to Spanish speakers and say okay, is this the right word for Spanish or not? Hey, can you translate this, crowdsource it, and double check it with everyone?

That way, you can not just translate 5 or 10 of the top languages in the world, you can be in every language all the time up-to-date using Facebook itself. Because oh by the way, when you translate the product way more people can use it. So translation itself, internationalization is a growth lever and we have product ways to do the translation with our users.

David: That’s what actually started even before the official growth team with Javi in that first hackathon. But that was the seed of this idea of, oh actually the way to get past this growth wall as people refer to it in retrospect that we’re hitting, is use the product itself. And internationalization, there’s even more to it than what you said.

One, when you have the users themselves who are translating a product as you’re seeding it in these new markets, they’re so much more bought in and feel a sense of ownership over it. Versus, oh I’m just airdropping this translated American product onto your country here. It’s not just that it’s a better translation, it’s that you’re actually seeding those first power users who are going to be deeply engaged and feel ownership over the product.

Then two, the nuance in this local markets that might even be in the same language, Spanish is the perfect example. Spanish is spoken in so many places around the world, but the local nuances and needs in Spain are super different than Mexico or Argentina or Latin America, et cetera.

Ben: That’s exactly right. I think my point with all this is: (1) growth is a pretty new discipline in our industry, (2) growth is not marketing. It’s very tempting. You see, especially with incumbents like big Fortune 500 companies who have someone whose title is growth and then you ask them what they do and you’re like, oh no, that’s not growth, that’s marketing, you don’t actually modify the product at all, that’s a different thing. It’s important to realize not only did Facebook invent the discipline, they are perhaps still the best at it. They really eat, sleep, and breathe the idea that growth comes from product.

David: It was that core initial team, all of them of which Javi’s now the COO, Alex is now the CMO of Meta, Naomi is now Meta’s longest tenured employee besides Mark and runs a ton of stuff at the company, and Chamath who deserves so much credit for starting the team, pulling it together, protecting it, advocating for it within the rest of product in the way that only Chamath can.

Ben: Man, it really is crazy with the Facebook diaspora. We’re going to keep seeing this as we go along. There are other names where people will be like, whoa, I didn’t realize they were at Facebook, but many of the names that have come up so far, yeah, the diaspora is pretty talented.

David: There were a million things that growth did over the years within the blue app and then ultimately within all the apps at Facebook. Growth today is a core center of excellence discipline that spreads across all the products at the company at Meta. But the other big thing in the early days besides internationalization is people you may know which crazy like the carousel of people you may know on Facebook is still a core part of the blue app to this day.

Ben: Dude, I just got lost in it. Did you know that that’s what I was looking at right now?

David: No.

Ben: I was sanity checking. I was like, oh is that person part of the original growth team? I opened Facebook, the newsfeed distracted me I scrolled down, there were stories at the top. I scrolled down one post, below that first post was an ad. Below that is people you may know. I am three panels over and people you may know because I was like, oh my gosh I’m not friends with them? We’re doing a podcast together live, and I just had that experience.

David: Totally. People you may know become such an important lever and I love it. I love it.

Ben: Okay, so people you may know become such an important lever.

David: For a whole bunch of reasons. The growth team discovers you were talking about analytics and data and really understanding what’s happening. The key thing that they discover for New Year’s, there’s joining Facebook about whether they become an engaged, active, evangelizing user or not, is how many friends they get in the first (I think) 24 hours or 48 hours or short period that they’re joining the network. And there are a certain number of active friends that you need.

Ben: Oh this is funny, by the way. I watched two different talks by people on the early growth team. One of them cited 10 friends in 14 days. The other cited 7 friends in 10 days of how do you create the magic moment?

Alex Schultz who’s now the CMO gives this great talk, gosh a decade ago at startup school with Sam Altman, and he makes the point. Look it’s a linear thing. Yes, you want as many friends as possible in the least amount of time. It’s not like oh magically at 10 friends in 14 days it’s super different than 9 friends in 14 days.

But you just set a threshold somewhere. You set the threshold and then you’re like, okay, if we can deliver this delightful experience where now people have a rich newsfeed and the people they care the most about in the world to interact with, they’re going to retain.

David: Totally. The other thing is it marries perfectly with all of the investment that is going into the algorithm automatization of newsfeed, because choosing who to show you in that carousel of people you may know is all the difference in the world. Show you seven randos and you’re going to churn. Show you your seven best friends and you’re going to friend them all, and become an engaged user.

Ben: I’m so excited for, I guarantee you at least one of the people that I just friended is going to be listening to this podcast and realize what just happened. Because I haven’t friended anyone on Facebook in years, and and I just sent out three or four friend requests.

David: They just reengaged you as a user. I love it.

Ben: They just have to, new growth tactic. Have people do podcasts about the company and then…

David: I think people caught onto that one a while ago. But yeah, that’s a super hard data problem of cold new user coming into the network. How are you going to accurately predict who their best friends are who are already on Facebook?

Ben: Well if you’re on the web and this is the late 2000s, you ask them to authenticate their web-based mail service of choice, find out who they’ve been emailing, and then use that to figure out who their friends are.

David: Magic.

Ben: There are other less nefarious ways, see who has sent them links in the past who are also logged in Facebook users. There are all sorts of stuff you can do, but yes.

David: So by the end of 2008 thanks to the growth team, growth is really going again. They go from just barely getting to 100 million users in August of 2008 to almost 150 million by the end of the year, so 50% growth in the last four months of the year. Then 2009 is just lights out. They grow 250% in 2009 to 350 million users globally. Internationalization is really firing on all cylinders here.

Ben: That’s right. They went from 150 million users to 350 million.

David: In one year. Isn’t that wild?

Ben: Crazy. The other core piece of the stat whenever you’re looking at growth is to look at engagement. Engagement in 2008 was also in the doldrums. Fifty percent of monthly actives were daily actives. From what I can tell that basically was just an artifact of as the company got bigger, every marginal user they added was less engaged than the early core users. When they went from colleges to high schools to open registration it, it just was going to have slightly less product market fit.

But the growth team is focused not just on growing new users. Actually, an even better lever, every long-term goal is retaining your existing users, and the best lever for retention is engagement. So that was where a huge amount of their energy went.

This is really interesting. Fifty percent of monthly actives were daily actives in 2008. Since then, they have basically improved that metric every year. There’s been a little bit of variance, but it is now almost 70% today. Been a maniacal focus on how do we make as many monthly actives, daily actives as possible year over year over year over year.

David: That’s wild. And that figure is just in the blue app in Facebook.

Ben: That is just in the blue app. Yup, exactly.

David: Wow. Well on the engagement front in 2009, the big thing that happens there is…

Ben: Like button?

David: You know where I’m going with this, like button. And I know you know who invented the like button, because it wasn’t Facebook.

Ben: I don’t know

David: FriendFeed.

Ben: Oh, no way.

David: Bret Taylor and Paul Buchheit invented the like button.

Ben: How did that work? Because FriendFeed would just federate stories from other social networks into one aggregated feed.

David: I don’t recall exactly.

Ben: Did it only store that information on FriendFeed and not propagate it back?

David: I suspect that’s probably it. They probably felt they needed a native mechanic on FriendFeed.

Ben: Man, FriendFeed was so awesome. It’s funny in this era where I’m now checking multiple feeds every day, Twitter, Threads, Instagram, all the messaging services to catch up on my messages, occasionally the blue app. I need FriendFeed again. That was the most amazing product.

David: I know. The world is so balkanized. It’ll never happen.

Ben: But yeah, talk about another amazing part of the Facebook diaspora. Facebook would later acquire FriendFeed. Bret Taylor would become…

David: CTO when Adam D’Angelo left to go start Quora, and then when Facebook Acquired FriendFeed, Bret became CTO.

Ben: And then famously Bret becomes the co-president of Salesforce by way of acquisition, and then eventually board chair at Twitter when the whole X thing was going on. And now he’s on the board at OpenAI.

David: Yup, board chair I think at OpenAI.

Ben: Crazy. Andy’s got a new startup of his own. I’m telling you, the talent that moved through that place.

David: Totally. Well and Paul Buchheit, the other FriendFeed founder…

Ben: Invented Gmail.

David: Created Gmail within Google, yeah.

Ben: So he ended up at Facebook for a while then after. I forgot about Paul, too. That’s unbelievable.

David: He must not have stayed very long if at all, because he went to YC I think after that.

Ben: Okay, so FriendFeed invents the like button. Facebook I believe was going to call it the awesome button. And then at some point right before launching, I think Mark personally made the call, it’s just too weird. Let’s go with like.

David: The story as I understand it is that Facebook had been considering something like this for a long time. The benefits are obvious. If you have a lightweight way to engage and feed more data about vote positive votes on posts into the newsfeed algorithm, that’s going to really turbocharge things.

Ben: On the flip side, the concern was that it would actually decrease meaningful engagement. There’d be less comments.

David: Exactly. Ultimately, they started running some tests with small user groups and specific geographies, and they found that it actually often boosted engagement of comments.

Ben: Fascinating. That’s why you test things. You never know.

David: Exactly. Yes they launched the like button and that becomes huge for overall engagement, but also more importantly for just improving the algorithm, which is now starting to kick in in newsfeed, and making sure you’re seeing stories that you really care about.

Ben: And the like button then also got used for pages. It became this form of capital among brands of how many Facebook likes does your company have? Which for a while actually meant you could organically get messages out to them. Every time you posted, it was as if a company was a person. You just show up in the feed before Facebook ultimately was like eh, those are advertisements. You’re going to pay us for those.

David: You’re exactly on the right track. Let’s flip back now to Sheryl and building the advertising business. Right around this same time we’ve decided, okay, yes, we’re going to be in the advertising business. To do that right, to do that natively, to do that in a way that generates defensible sustainable value for Facebook, that means targeting, and the like button is the perfect gift to this. Because all of a sudden, there’s this really lightweight way to engage with things, to signal your preferences about what you like.

Ben: And the like button ended up being the perfect Trojan horse to move Facebook platform off of Facebook. There was a reason for third-party websites to embed Facebook’s JavaScript in their pages, because of course you’d want to be able to like an article or like a brand all up on their website. How many Facebook likes do you have? If that’s what matters, we want people clicking like on our website too, not just over on Facebook.

Suddenly, now every company on earth has some Facebook and has a reason to embed some Facebook JavaScript right there on their page, which my goodness, that’s going to serve as a great signal back to the advertising algorithm, where we can drop cookies, and we can see who is moving around the web in what ways. It’s perfect for platform, but it’s also perfect for feeding data into their advertising system now that they have something that brands and third-party websites are incentivized to drop right on their page.

David: Abso-freaking-lutely. We talked about growth reigniting at the end of 2008 into 2009. The business is also on fire in 2009 in a good way. Platform still rocking and rolling. Microsoft display ad partnership still growing strong as growth is happening. Now under Sheryl, you finally got real, native, highly-targeted ads coming online in Facebook.

In 2008 the company had done $280 million in revenue, up from $153 million the year before in 2007. 2009, they do almost $800 million in revenue. He’s crushing it. 2010 an even $2 billion in revenue.

Ben: They’re now very, very cash flow–positive.

David: Reaching the scale that Google was at their IPO. 2010, they hit 600 million users, 2011, 850 million users, 2011 revenue grows almost another 100%, $3.7 billion. All right, time to start thinking about taking this company public.

Ben: It’s also time to start thinking about what is our biggest existential threat. It’s so funny that we’re talking about and the existential threats are over. We’ve won. We have platform going well, we have this advertising business going well, we’re getting data from the open web because we have like buttons everywhere. Everything is going our way. We finally have it dialed.

And none of that matters in a mobile world. As people shift from spending time on the web to other apps, that open web data used as signal goes away. All of your ability to take payments. By the way, Facebook had launched payments. At some point, they started requiring apps to use their virtual currency.

David: Facebook credits.

Ben: Facebook credits, that’s right. All of these amazing pillars that they’ve built were for the open web and created the most incredible business known to man, and mobile is a completely different thing.

David: That seed from five years ago that has been growing on mobile is finally going to turn into a big, big, big problem for the company.

All right, so let’s just get into it. I thought it would be fun. I listed out all of the ways that I could think of that mobile was a big problem for Facebook. One and I think this is like the Uber problem that everything else is downstream of, iOS and Android are closed, proprietary ecosystems not controlled by Facebook. Yeah, you can argue Android is open source and all that, but for all intents and purposes Google controls it and they control the Play Store.

Ben: Put another way, the web is the only open platform in history and Facebook was born on the web. How crazy is it that they could build Facebook on entirely free technologies at the beginning, and then get distribution just by people sharing URLs around.

Browsers are interchangeable, operating systems are interchangeable. It works on any device that anyone wants. Yeah. If you live on the web, you have infinite degrees of freedom and flexibility to control your own destiny.

David: And Facebook had built this walled garden on the web, and it was pretty nice.

Ben: And walled gardens are great as long as the foundation under you can’t shift. Build your walled garden directly on the earth, not on someone else’s foundation.

David: But okay, specifically why is iOS and Android being closed, proprietary ecosystem such a problem? That means my number two reason. Platform which is 50% or more of the value of Facebook in their minds at this point in time, no longer works at all. Full stop, flip a switch, does not work, will never work at all ever again on these platforms.

Ben: And why is that?

David: The whole premise of platform is you can run apps inside of Facebook. On mobile, you will never be able to run an app inside another app. Just not going to happen. Apple’s never going to let it happen and Google’s never going to let it happen.

Ben: And that’s primarily because of that thing around you can’t launch an alternative app store.

David: Yeah. It’s the core part of their business model are the app stores.

Ben: Not to mention, if you are paying for things on their phone, you have to use their payment system. Facebook credits ain’t going to work because you can’t charge 30% on top of 30%.

David: And there are technical aspects to this too. Code runs in native code on these devices as opposed to open HTML on the web,

Ben: You get told what your development environment is. This is the language, these are the frameworks, these are the APIs you’re allowed to call.

David: Okay, greater than 50% of the revenue and the…

Ben: Theoretical future value.

David: Inherent theoretical value and strategy of the business, kneecapped. It gets worse. Facebook and Mark and Sheryl have been hard at work building the other core pillar of the business, this new, super sexy targeted ads business. Well guess what? All of the ads that are running on Facebook at this point in time are running on the right hand column on the website. There is no right hand column on a mobile app. There’s only room for a feed, which to this point in time has never had advertising in it.

Ben: And in fact, there a cultural allergy to the idea of polluting the beautiful, pristine, organic newsfeed with an ad. It will require big cultural change.

David: This is why you can’t make up this stuff in these stories that we tell. Facebook has, I kid you not, no way to make money at all on mobile. Here is an actual honest to God statement that is in Facebook’s S-1 that they filed before the IPO in 2012.

I am reading as a quote, “We had more than 425 million monthly active users who used Facebook mobile products in December, 2011. We anticipate that the rate of growth in mobile users will continue to exceed the growth rate of our overall MAUs for the foreseeable future, and our users could decide to increasingly access our products primarily through mobile devices.”

Ben: Which let’s be clear, comparing those growth rates, another way to simplify that and say, it is our user base is shifting to mobile from desktop.

David: Another way to simplify that is, our desktop business is going to zero. Now the kicker. This is a statement in the S-1, “We do not currently directly generate any meaningful revenue from the use of Facebook mobile products, and our ability to do so successfully is unproven. Accordingly,…”

Ben: Buy our stock.

David: Yeah. “Accordingly, if users continue to increasingly access Facebook mobile products as a substitute for access through personal computers, and if we are unable to successfully implement monetization strategies for our mobile users, our revenue and financial results may be negatively impacted.” Yeah, no kidding. Oh my god. I’m still not done. That’s only reason number three.

Okay. Reason number four, another benefit of the open web is that you can push code whenever you want. You can make changes rapidly. You can move fast, you can break things, and then you can fix them rapidly. This is how Facebook has always operated. On mobile apps, you can push code when Apple and Google tell you that you can push code.

Ben: And at this point in time, there was approximately a two-week delay between pushing an update and it being reviewed and accepted.

David: Not only is there a two-week delay, during that two weeks Apple and Google are reviewing, if they see anything that they don’t like for whatever reason, they are telling you no you cannot push that code.

Ben: Or your company could be in a fight with that company and they could decide, eh, I don’t think you should push any updates for a while until we resolve our fight.

David: Yup. Okay, that’s number four. Then that brings me to number five reason that I could think of of why this is really bad for Facebook. The competitive field is about to get a total reset.

On the web, Facebook is dominant. Network effects are super real. Everything we talked about in this whole first part of the episode, if they can suck up all of the new social functionality that they or anyone else invents, they can put it into the main Facebook app—photos, chat, platform, newsfeed, et cetera.

On mobile though, Facebook is just one icon of many on the home screen.

Ben: Exactly. The wisdom at the time was that mobile apps should be narrow in their functionality and you did not expect a single app to bundle in a ton of different use cases.

David: It’s super easy for users to be like, hmm, okay. I get this utility from Facebook. But well, I don’t know. There’s this Instagram thing. That’s cool. I get great photo utility from that, which also lives on my home screen right there. I’ll still use Facebook for seeing when my friends get married. But photo’s cool. And messaging, you know this WhatsApp thing is super cool. Maybe I’ll just use this app for messaging. It’s super easy to use one app for one thing and other apps for others.

Ben: And you can see why Facebook adopted that early 2010s constellation of apps strategy. For a while, they had Slingshot, Poke, Messenger, Paper, Rooms, Riff…

David: Camera.

Ben: The belief by a lot of companies for the direction mobile was going was there are going to be specialized apps each for their own tiny little purpose. That’s not great if a lot of your value is we bundle a lot of stuff in to create the most user user engagement to all feed into each other for people to use all these different components of our application.

David: Totally. Okay. Those are my short list of five reasons of why the shift to mobile is dangerous to Facebook. Anything else you would add?

Ben: No. So they’re going public right into this. For the first time since our business is founded, we face a real existential threat completely out of our control that is going to make the next few years look really bad. Let’s go public.

David: Let’s talk about that IPO.

Ben: Yes, but first it is time to talk about one of our favorite companies Statsig. A phrase that many of you will know from Facebook’s early days is move fast and break things. But despite instilling this in Facebook’s engineering culture, Facebook didn’t actually break very often, and it essentially never goes down now. How?

David: Well, Facebook invested hundreds of thousands of engineering hours in a set of internal tools. These tools let any engineer set up new metrics, ship new features, and measure performance in real time. This meant that anyone could just ship a new feature.

But they always had metrics to use as guardrails, and they could always roll back the feature if anything broke. They took this so far that they have every new engineer ship a feature on their first day at the company.

Ben: Wild. You might wish your team could build products like Facebook. Ship fast, make data informed decisions, iterate rapidly. But you don’t work at Facebook, and you don’t have those tools, so you’re stuck, right?

David: Well, not anymore thanks to Statsig. Statsig was founded by an ex-Meta team three years ago with a mission to make these same tools that Facebook has available to any company.

Ben: Could there be a better sponsor for this episode, David? This is insane.

David: I know. It’s perfect. Today, they’ve done it, combining tools like feature flags, product analytics, experimentation, and observability into one connected platform that runs off one set of data and infrastructure. No more stringing together expensive point solutions and internal builds, and no more using technology vendors built 5 or 10 years ago when the needs of modern product teams have grown by leaps and bounds.

Ben: And they’ve also gotten some crazy traction over there at Statsig. Many of the world’s leading tech companies rely on them, including OpenAI, Microsoft, Notion, Anthropic, Figma, plus thousands of early-stage startups. In fact, their scale has gotten so crazy that they process over a trillion events per day. For any engineers listening, they have a great blog post about how they do this.

David: And best of all, Statsig is pretty affordable. They have an insanely generous free tier for small companies, a startup program with one billion free events which is $50,000 in value, and significant discounts for enterprise customers. To get started, just go to statsig.com/acquired and remember to tell them that Ben and David sent you.

Ben: Thanks Statsig. All right, so David, here we are, we’re going public despite everything you just listed that is wrong with the future of mobile and this company.

David: Oh man, which seems like a not very smart idea. One thing about Mark Zuckerberg, you could say lots of things about him, but one thing I don’t think anybody could say is that he’s not smart. What gives? Why are they doing this?

Well, back then, once a company crossed 500 shareholders, the SEC actually mandated by law that you had to start reporting quarterly like a public company.

Ben: Whether or not you were traded on an exchange.

David: Whether or not you were traded on an exchange. At the end of 2011, Facebook crossed 500 shareholders. For Facebook, this means that they were going to have to report like a public company anyway.

In practice, the common wisdom was you have to go public as part of this anyway because otherwise, you’re going to report. Then you’re going to lose all the momentum and excitement about having an IPO, doing an S-1, having a big road show, et cetera. You might as well just do the IPO as long as you’re going to have to do this.

I think Mark, Sheryl, and the team really debated, well should we skip the IPO? But well shoot, then we’re going to be reporting as a public company and heading into all of these headwinds without having actually gone public and raised the money that we might need to be able to make these investments. They’re like, nope, we got to just do it and go public.

Ben: Yes, it is an odd time to go public, given everything with mobile, and of course, they’re being forced into it. There actually are a couple of tailwinds that they have that are probably worth touching on here. The first of which is they had just beat Google+.

David: That’s right.

Ben: Facebook treated this like a total existential threat. We laugh about Google+ now it’s a butt of jokes, but that’s because Facebook was so effective in competing against it. I’m not saying the product itself was amazing and it would’ve been fine without Facebook, but Google did not end up executing that well on that product.

David: It actually is a really good point though. Google+ may or may not have been a good product and may or may not have succeeded, but the Facebook competitive response was unbelievable.

Ben: This is the growth team plus Facebook’s engineering culture at its finest in defeating this. Whether or not they executed the product well, Google was the big web tech company at the time, and they put a ton of resources and a lot of their best people on it.

David: And they had Gmail.

Ben: And they had Gmail, right. It’s interesting to look at this because Google basically is Facebook’s biggest business model comp. They show ads on the web and they monetize that really well. This theoretically could have been in their wheelhouse. This is just more real estate to show ads on the web, and they’ve already got all these people with user accounts. You can see why Facebook took this really seriously.

We’ve talked a lot about Facebook’s technical prowess. Well, here is an example of where it really kicks in when you need it to, to be a key competitive advantage. They structured themselves in a way that encouraged flexibility of engineers, and they really broke from the tide of microservices in this era. They had one monolithic code base that everyone worked out of. You might say, well that’s stupid. Why would they? That’s not the way the world was going.

But what they basically did was they wanted to encourage portability of teams. If you’re an engineer, you get hired into the company, not onto a team. You learn the company’s code base, you learn the company’s conventions, and you can easily move around after that.

You do have to deal with the trade-off then of you have this big monolithic code base with gigabytes of PHP code for thousands of engineers. What do you do about that? Well then they just had their cake and eat it too. They would go and have the infrastructure team figure out how to deal with that problem rather than saying, oh, everyone can just work in their little silos.

That meant that engineers could be quickly reorged. They could have this company-wide lockdown to fight Google+ and they did all sorts of things. They launched video calls to compete with Hangouts. They launched something to compete with Google Circles. But either way, they could really quickly reallocate resources and people who knew how to work together to defeat what could have really been an existential threat from their biggest similar company.

David: Oh man, Mark famously in an internal speech around this referenced, what was it? I think Cato the Elder? From Ancient Rome of Carthage must be destroyed.

Ben: Yup. There is a second way in which things had settled down and now was a good time to go public, and that’s around privacy. Facebook had just been playing way too fast and loose with user privacy for years and it had finally caught up with them by 2011.

Just to jog your memories—I’m sure people may remember a lot of these, some of these included—even though users could restrict the audience of their posts with a setting, this apparently didn’t apply to apps, which could access these posts regardless of how you restricted the audience. For a time, this even included when your friends installed an app too. You didn’t even have to grant the permission yourselves.

Similarly, they made friend lists public without user consent, at one point. Facebook could decide without user consent to change private fields of your profile to be public. This wasn’t always messaged as clearly as it could have been to users.

To remedy this, they had just signed what is called a consent decree with the FTC, the US Federal Trade Commission in 2011. They promised to make a bunch of changes regarding user privacy issues going forward. So all this was behind them now and interestingly David, an FTC consent decree is the same thing that Microsoft signed.

David: Hard back to the Microsoft episode.

Ben: The FTC consent decree with Microsoft was the predecessor to the big DOJ suit. In this case, the FTC consent decree is the predecessor to what eventually would become the Cambridge Analytica settlement.

Here in 2011, the way that they settled this is there are a bunch of provisions that with Facebook promising they’ll be tighter on making sure user data is treated in a very particular way. They’re subject to audit every year for two years, for the next 20 years. There are all these things that they agree to.

But once you sign a consent decree it’s like okay, we’re through it. We don’t have this thing hanging over our heads. We can go public and say, oh yeah, that’s in the past, we’ve taken care of it. That’s another reason to go public right now.

David: It’s such a good point. There actually was this little window here at the beginning of 2012 where desktop revenue hadn’t yet been impaired so badly that they could still show growth. They had these couple of little wins they could point to, and say, all right, not great but if we’re going to do it, we got to do it now.

Ben: You could almost look at it like a little bit of a win, of everyone knows us as the company that is a little bit dodgy on privacy practices. At least the public perception is this, that they’re constantly…

David: Changing the terms of the game, shall we say.

Ben: Yes, to the company’s advantage when it was confusing or misleading to users, And now you could say, look, not only did we sign that. We have these five product initiatives in place where we’re just buttoned up now. I actually think that’s pretty true. I think they became a company that had rigorous privacy practices because of some government regulation when they otherwise may not have.

If you look at the early days of what you could do as a developer on platform, it was pretty wild west. I’m not saying it’s fully because of the consent decree, but they could definitely tell a story around, look, we made some mistakes, we got them some things wrong in privacy, and going forward, we’re in good shape,

David: We’re all good, yeah, Well, we’ll come back to that in a little bit. But for the moment, the IPO. February 1st, 2012, they filed the S-1. They entered the traditional quiet period before the IPO where legally, according to the SEC, once you have filed for an IPO, all material information about the company needs to be included in the S-1 registration statement.

You can’t go give interviews, you can’t go talk to the press, you can’t say anything meaningful about the company besides what is fully publicly disclosed to any potential investor in the S-1 statement. Then a week later on April 9th, 2012, during the quiet period, Facebook acquires Instagram for $1 billion.

Ben: Time’s now baby.

David: The time is now. They can’t say anything about it besides what they file in the amended S-1 because they’re in the quiet period.

Ben: Amazing.

David: The company’s previous largest acquisition was $70 million. Obviously, this is a way different thing. Now sitting here today, you might be really tempted to say, oh well there’s the fix for mobile. They just bought Instagram. That solved all the problems. Except not, because Instagram didn’t have any revenue on mobile either. You’re tying one anchor to another here.

Ben: Well not to mention Instagram had 27 million users, Facebook had 900 million users. This was potentially a problem for future Facebook, but this was not currently a problem.

David: Totally. I was just about to say that too. Also not currently a fix to Facebook’s major problem of the shift to mobile. My God, talk about the chutzpah of Mark to do this. You’re already in a really challenging period for the company. A really challenging IPO process. You just entered the quiet period, and you decide to spend a billion dollars on something with 27 million users and no revenue, and you’re not going to be able to talk about it except what you file with the SEC.

Wow. You got a really, really, really freaking belief in the future potential of your company, and the combined companies of Facebook and Instagram here to make this move.

Ben: Not to mention on top of it all, we’re not going to go into it here because we had a whole episode on Facebook and Instagram, but this was done over two days over the course of a weekend. Mark didn’t do a whole lot of socializing this before pulling the trigger. He just knew it was the right thing to do and did it.

David: And that ended up working out pretty well in the long run in the short run. Yet another question mark to add to the already very large pile around this company. Finally, long roadshow process. By the week of May 14th, they’re ready to price the IPO. Pricing is set for Thursday night, May 17th, 2012. On Tuesday, May 15th, General Motors (GM), who by the way was not a particularly large Facebook advertiser, announces publicly to the world that they are pulling all their spend off of the Facebook platform because it’s not effective. Not really what you want to have happen two nights before you price your IPO.

Ben: Was that motivated by someone? That feels like a hit?

David: It feels like it must have been. I’ve never heard one way or the other, but: (a) why would you announce that publicly? (b) why would you announce it publicly right then? There’s for sure more to that story.

Nonetheless, Facebook goes ahead. Thursday night, they price the IPO at $38 a share at the very top end of the range after the road show. Again, the chutzpah here. They raise $16 billion at a $104 billion market cap, making it (I think) the third largest IPO of all time at that point behind Visa, and ironically General Motors coming out of the bankruptcy during 2008.

Ben: Which both of those are mature companies that are going public under weird circumstances. Those are not comps.

David: Totally. By far the largest IPO of a new startup company in history. On Friday morning, trading starts. The NASDAQ is overloaded with so many orders for buying and selling Facebook. That trading and Facebook stock doesn’t actually open until 11:30 AM, which is, two-plus hours after the open of the markets

Ben: got to be terrifying if you’re a big Facebook shareholder.

David: Yeah, that’s pretty terrifying. Initially it looks like, oh that delay might have been good because the price shoots up to $45 a share from $38. But then it crashes back down. It is one of the most volatile trading days in any stock, I think in history on any exchange.

Facebook shares end the day at $38.23 up 23¢ from the IPO price. But only because all of the underwriting banks in the IPO stepped in to buy shares and support the price, which was totally crashing throughout the day. Thank God They did it on a Friday, so everybody had the weekend to breathe.

Once trading starts again the following week, the price falls on 9 of the next 13 trading days, and ends up by the end of May down 25% from the initial IPO price of $38. In free fall at this point, the stock continues going lower and lower and lower. Ultimately, bottoms out on September 4th, 2012 at $17.68, down 53.5% from the IPO price. Market cap cut in, over half in, what’s that? Five months since the IPO? It wouldn’t end up reaching IPO price again until August, 2013.

Ben: Anyone who bought at the IPO was underwater for 16 months.

David: Underwater for 16 months, and 5 months later had lost half their money on Facebook.

Ben: You have to have nerds of steel to even just keep holding. I’m sure anybody who got back to even at 16 months was delighted.

David: Totally. You’ve also got the employee lockup coming up around November. Now you’re staring down the gun of, shoot, I’ve already lost half of my value as a public company, and pretty soon all the employees in the company are going to be able to sell it. That’s going to put more selling pressure on the stock. Holy crap, this is pretty darn bad.

Ben: Yes, that is horrible.

David: Losing 50% of your value as any company, public or private, but especially a public company, is a disaster of epic proportions. Having that happen immediately following your IPO for most companies, there’s no way to come back from that. You’re entering a death spiral, you’re going to end up being acquired or you know you’re just going to get driven to zero basically.

And just to be super clear, desktop revenue wasn’t going to go to zero immediately, but Wall Street and investors are really worried that like hey, the world is transitioning to mobile. Eventually growth will start slowing and it will start declining, which it does fairly quickly here. If you can’t show me that you’re going to replace that with mobile revenue, why should I value your stock at anything?

Ben: So here is where founder control matters. David, it’s interesting you’re explaining the mechanics of a death spiral. That would presume there’s a board of directors who feels a strong desire to do whatever is best in the interest of the shareholder. They might think that it’s these short term things.

But if you have a CEO who has a 20-year-view and believes deep, deep in his heart that the best is yet to come, and that this is just a temporary problem, and that person controls the company and owns the majority of voting shares, you can withstand things like this, you just better be right.

A little fun anecdote for everyone. David and I, among the 20 other people we talked to to prep for this episode, one of them was Sheryl Sandberg. We were asking Sheryl, in particular, how did you start the monetization effort on mobile? They were in the third column, there’s no third column on mobile, what did you do?

Her comment to us was, oh, we just stopped caring about the right side ads on desktop. We took every engineering resource we could off of that, and we put it on mobile. We knew we were going to miss the current quarter—I think they missed a lot of quarters right after their IPO—but this was us trading the present for the future, and all we cared about was our future. She said, this is great quote.

David: This us up there with one of the best quotes on Acquired of all time.

Ben: She was sitting there with Mark late at night, and when they arrived at this plan of we are going to forego a lot of desktop revenue to basically bet it all on figuring out mobile revenue, she said, “Well Mark, nobody can fire you, and only you can fire me. If you’re in, I’m in. We buckled our seatbelt and we said, ‘here we go.’”

It’s amazing. That can only happen in a founder-controlled company, and it really did force them to figure out mobile advertising. If they really are saying, this is the new first class product, this is where we’re going to point all advertisers toward, if they’re wrong on that, they’re wrong across the whole board, because if ads suck on mobile, since it’s the only thing and it’s filling up your whole screen, user engagement is also going to nose dive.

It’s basically a be at all moment where the ads are front and center, so they must be good. It is essential for the mobile product and thus your company for them to be good. User’s attention will be pointed at these ads like no other ads you have ever run before.

Actually, the flip side of this is it ended up being the best thing ever for the company. Because the ads are front and center, the value per ad is actually higher. They ended up creating a much, much more valuable ad unit than they ever had on desktop purely because of this incredible attention on them when you’re scrolling in feed on mobile. Necessity is the absolute mother of invention.

David: Totally. You have no choice but to do your best work, and boy did it work out fabulously. Ben, nobody can fire you. And I guess you can’t fire me either. But either way, if you’re in, I’m in.

Ben: The corporate structure of Acquired is still a little unclear.

David: But yes, totally. When we say taking your best, most trusted engineers, taking them off whatever they’re doing and putting them on ads, there’s actually an amazing story around this.

Mark goes to someone within the company who he knows can just ship and get stuff done. That is Boz, Andrew Bosworth, his old TA from Harvard. Boz at this point is running the core profile and timeline within Facebook.

Mark goes and talks to him and says, Boz, I think you should come work on ads. To which Boz replies, okay, I get we’re in this situation, but you know Mark, I can’t do this for too long because remember I’m getting married this fall, and I’ve got this six month sabbatical plan for my honeymoon. My wife and I are going to travel the world. It’s all booked and paid for. We’re going to be out of the country. Mark is like, yes. Well we’ve got a few months until then. I think you should come over and work on ads.

Boz was like, okay, okay. I will do that. I will come up with something here before the wedding. He teams up with Will Cathcart from the newsfeed team, from Chris Cox’s team, and they go get to work. By the way, Will is the head of WhatsApp today and Boz of course is the CTO of Meta and head of reality labs.

The first thing they do during these couple of months is like, all right, we just need a bandaid to stop the bleeding and at least produce some mobile revenue. What they decide to do is like, hey, we do have on mobile the carousel of people you may like. What if we just slot some brands into that carousel and say, hey, in addition to people you may like here, here are some brand pages you may like.

Ben: A reasonable first step.

David: And look, they knew that this wasn’t going to be the long-term solution and solve all the problems, but it did start to produce some mobile revenue that they could report. This was super important because as I alluded to the lockup on employees and insiders selling their shares expired in November. They had to get something out that was going to prop the stock price up a little bit before naturally employees are going to start selling here and that’s going to put downward pressure on the stock price.

Ben: Also, bummer if you are an employee selling six months after that IPO. Hopefully, as many people held as they could to get through it.

David: Yes, hopefully. I think most people who were at Facebook at this time really did believe in the future of the company. At the same time, you have life circumstances. You might need to buy a house, you might need to do things for your family, et cetera. It’s a tough spot.

Fall rolls around, they’ve launched sponsored pages you may like. Boz gets married, they’re still grinding on the big project, which is called Project Whale Shark. That is getting real honest-to-God native ads in the newsfeed, which frankly they should have done years ago on desktop. Now, the gun is to their head. There is nothing in the mobile app except newsfeed. We got to put ads in newsfeed.

Finally, at the end of the year they launch it. Boz is like, whew, okay, mission accomplished. I can now leave on my sabbatical, we can go on our honeymoon. But there’s just one problem. Shipping this thing is only 20% of the battle. Now you have to go sell it to advertisers, which this is a whole new ad paradigm. You got to like educate brands on why this is going to work for them, especially when your overall brand as a company and effectiveness of your advertising is in the dumps right now.

Ben: And it’s an iterative process where you’re going to hear a bunch of feedback from brands and you’re going to say, ooh, we got to modify the ad products, which still continues to this day. They roll out and modify ad products. This is not all right, we now have ads in the newsfeed. We’re good.

David: The job is never done here. Of course, quarterbacking all this is going to fall to Sheryl and her team. But again, as you said Ben, this is an iterative process. You can’t divorce product from the feedback from advertisers. And the big advertisers are sure as hell going to expect someone from product who is leading this thing to show up and tell them about it. So Boz needs to come along to these meetings.

On December 18th—Boz remembers the exact day—Mark calls him up and is like, Boz, great job working on ads. Hey, I really still need you to come back and work on ads. Boz is like, no way man. I did the past six months for you. I just got married. I’m going on my honeymoon. I’m out. Mark’s like, okay, okay, okay.

The next day calls him back December 19th and he’s like, Boz, I just want to say really, really great job again. The company’s in a really tough spot and I really think that you should come keep running ads here. Boz is like, okay, all right. I will reduce my sabbatical from six months to six weeks. I’ll take the holidays the beginning of the year. I’m going to cancel the last four months of my trip. I will do six weeks with my wife and then we’ll fly her friends out to the last two weeks with her. Then I’m going to go fly around the world with Sheryl on a very different trip here. That’s what they do.

For most of 2013, Boz and Sheryl are out on the road. They’re pitching advertisers, they’re explaining sponsored stories in the newsfeed on mobile. Why this is so important. Not only do advertisers need education about this new Facebook ad unit, they also need education about mobile advertising in general. They’d only just gotten used to digital advertising on desktop.

Ben: Totally. Do you remember those Mary Meeker decks that used to go around? Where it would show the mobile advertising, basically the shortcoming. Look at all this attention that has shifted to mobile and yet the monetization just isn’t there. That was a story every year for a decade.

David: Totally. It was the same story with desktop on the web for a decade before that. What’s interesting is as they’re getting going here, of course advertisers don’t come in right away, but also the ones that do come in early are kind of, as you would imagine, not the best advertisers.

So the whole effort starts to have this real problem where the ads that are now showing up in newsfeed are pretty freaking crappy ads. Mark keeps getting more and more pissed, because obviously this is really important. But he’s like, Boz, Sheryl, why the hell is all this crap showing up in my feed? If this keeps happening, I get that we have to save the business, but I care more about the user experience and I’m going to dial this back.

Ben: This is the most interesting thing. At this point in history, Mark is putting pressure on Boz and Sheryl to show fewer ads. And Boz is the one with the contrarian view who comes to Mark and says, we’re thinking about this all wrong. We actually need way more ads. Not just a little more ads, huge amounts more ads, because a great ad is on par with content. If you have tons and tons of ads, then we can do way better targeting.

You have this beautiful liquid marketplace of hundreds of thousands of advertisers, billions of users, and at any given time, somebody can see the optimal ad for them and get recommended an amazing product that perfectly fits their needs in that moment with messaging that’s perfectly tailored for them.

Great advertising can be great, but you need a really, really liquid marketplace, and you need fantastic algorithms, which you can only have if you train them on tons and tons of data. We actually need 10,000 times more ads than we have right now, not less. It’s almost like the only way out is through. We’re going down this path. We must be extremely successful in order to be successful at all.

David: This was such a contrarian view and also right. I think only could have come from someone like Boz being focused on this problem, because as a product person and an engineer, it’s the only way that you’re going to have that insight that like, oh, this is the same problem we had when we launched newsfeed. The way to fix newsfeed and to make it really, really awesome is more newsfeed, is more data.

Ben: The only way out is through.

David: Yeah. If somebody had purely come (I think) from the advertising world, they would’ve been like, well, that’s just the nature of advertising. It attacks on the users and that’s the way monetization of media works. But Boz was like, no, no, no, actually this is a relevance problem and we can make this awesome. In order to make it awesome, we need a lot more of it.

Ben: And basically like that’s what happened. Flash forward 10 years, I open Instagram to get great product recommendations.

David: Totally. It’s actually wild. I believe this is true. There are markets in the world today where I think Meta has run this test, where if you turn off ads in the Meta family of apps products, engagement in those products actually goes down, which is just freaking wild. You would think, oh wow. Now I get to use Meta products without advertising. No tax on me. I’m going to engage more. No, it goes down.

Ben: That’s the definition of ads as content, if that’s actually the case.

David: Totally wild. By the fourth quarter of 2013, which honestly in the scheme of things is pretty freaking fast, the whole thing finally flips. Facebook announces that mobile advertising is now 53% of revenue, and total ad revenue for the company grew 76% year over year in the fourth quarter, which of course is the most important quarter of the year.

I did all the math on that. That means that all of desktop revenue, all of it—the old platform business, the old ad business, the Microsoft third party ad serving—all of that actually shrank in the fourth quarter of 2013. That’s the counterfactual here. If Facebook had not gotten mobile monetization right, that’s what would’ve happened is Q4 2013, all of the growth momentum that it had would’ve flipped and it would’ve started the downward spiral.

Ben: And probably not that early if…

David: I guess that’s right because all the resources got pulled off of it.

Ben: Exactly, but that’s what would’ve happened eventually to it.

David: Eventually, yup. Man, it’s this wild 18 month journey, but it totally saved the company because now here they are, they’ve built this unique pioneering and totally defensible native mobile ad unit in the feed. They define the whole generation of monetization here.

Ben: It’s the best ad unit in history. It’s an ad that completely fills the screen on your device and that users are okay with. This is the most insanely captivating, engaging ad unit that you could possibly ask for. And it came out of necessity. It’s wild.

The fact that they thought their backs were against the wall, they were totally screwed. instead, actually it’s the thing that monetizes better and is better for advertisers than any other advertising product they’ve ever had.

David: I know we keep saying this, and it sounds maybe silly, but really better for users too. Before this paradigm that they invented here of algorithmically-driven ads in the feed, the biggest monetizing thing on mobile was an interstitial popup fricking display ad that took over your screen. What an awful fricking experience for a user.

Ben: Right around this time, too, the other thing that’s happening is—kind of glossed over this for time—Facebook’s initial attempt at a mobile app was to try to work around all the constraints of the mobile app ecosystem. They thought, well that’s nice that the app stores are going to try and box us in, but we’ll just ship our mobile web views inside of a thin little app wrapper, and that way we can deploy multiple times a day. We’re Facebook. This is what we do. It’s part of our culture.

David: HTML5, whoo.

Ben: Provided a horrible user experience. The engagement on the app was low, time spent was low. It was a risk to start selling these ads because people don’t want to spend any time in the app even without ads, let alone when you start layering these in. So they’re finally starting to pull out of this tailspin. They basically spent a year completely rewriting all their mobile clients to be these rich, beautiful, native experiences.

This is a thing that Facebook has always been good at. Whenever they decide to do something, they go and recruit like the actual best talent in the world to do it. This group of people that they pulled together from Joe Hewitt forward to write their original iOS app. It’s just some of the best iOS developers and designers in the world. They hired Mike Mattis and the Push Pop Press team when they acquired them.

David: That’s right.

Ben: That became, I think it was Creative Labs, is that right? Facebook Creative Labs. But a huge amount of that talent worked on their mobile apps. So while they had had the wrong strategy at first, once they got religion around native, they really created probably one of the best apps ever on mobile.

David: Totally. It’s funny, I’m laughing. The version of history that we just told is maybe slightly too charitable to Facebook and to Mark because yes, it’s absolutely true. They did the really hard thing. They took the pain, they invested in the future, they built everything that would become Facebook and Meta today out of it. But the first reaction was to bet on hope as a strategy and bury your head in the sand and say HTML5 is going to save us.

Ben: I don’t know that it was hope as a strategy. I think it was more like if this can work, it’s going to solve a lot of our problems. I don’t think they correctly estimated how wide the user experience chasm is between web apps and native apps. I think they had to have an app in market where users were actually using it to realize, oh man, the state of the art in native that the platform vendors have developed (iOS and Android) is really, really good.

They have brought very little of that to the web experiences, partially because of standards bodies, but also partially because it’s not really in their interest to make web apps great when they can force everything through an app store that they have more control over.

David: After the Q4 2013 earnings announcement, the stock pops way, way, way up to $175 billion-ish market cap, so up 75% from IPO, what, is that year-and-a-half earlier?

From there it’s basically, if you look at the market cap chart, a straight line journey from 175 up to half a trillion over the next couple of years, they eventually bring this whole native feed monetization engine over to Instagram. That works obviously incredibly well there, probably better than even on the blue app.

Ben: And that, in particular, everyone’s going to laugh when we say this word. The legitimate synergy between going to an advertiser and saying you can use this dashboard to get placements on Facebook and on Instagram, is massive. Both of those products monetize better than they ever could without that single channel that the advertiser only has to go through and use one dashboard to place on both products.

If you flash all the way forward to today, the lion’s share of Meta’s revenue comes from the ads that run on Facebook and Instagram. Their whole business today can just be summarized as that.

David: Totally. There’s another element for a couple of years that’s a big part of the native mobile revenue story, and that is app install ads. Actually, is worth calling out because it plants the seeds of this much larger struggle that’s going to come in a little bit.

But a very large percentage of the inventory that they were selling and especially the revenue of these native mobile ads in the beginning was actually for app install ads that game developers and other folks were buying on Facebook. It was like our old platform business all over again. Facebook is actually the best place to discover apps on mobile.

Ben: It’s a pretty interesting realization where the big brand advertisers may not move to this new type of ad format right away, but the people that are going to be really hungry to move to that ad format are a game developer who makes a mobile app and wants to market their mobile app to people who are: (a) on that platform, and (b) in a lean back experience where they’re open for some entertainment.

When you are scrolling through a feed of your friends and brands, and you are open to, oh hey look at this, a game where I could click one button and then boom, install a game and play it, is there a better moment and channel to reach someone for an app? No. Even if you’re Apple, Apple doesn’t have a better way to do this.

People don’t search the app store for apps that often, so you’d have to show them a pop-up ad or something. Facebook just has this opportunity where you’re in an experience where you’re open to some new form of entertainment, and they have the ability to place a button there with rigorous targeting and an incredible ad sales force. Facebook was almost built to be the monetizable front end to the app store.

David: Oh boy. All right. Put a pin in that.

Ben: Am I leading the witness too much here?

David: Put a pin in that. We are going to talk about Apple and Facebook in just a little bit here.

Now I said it a minute ago that the journey to $500 billion in market cap from here is a straight line. If you look at the market cap chart that’s true, internally within the company, I think almost certainly not there are a couple of big challenges that come up along the way. During this period, you’ve got Snap and you’ve got WhatsApp, both of which are pretty big deals. We’ve covered both of those pretty fully on separate episodes that you can go listen to.

Obviously, Snap was more a competitor and WhatsApp turned into an acquisition and is now a big part of Meta and part of the company. But I think together what they represent was this next fundamental shift in the consumer social landscape, which Mark totally picked up on and ultimately announced as the core of the company’s product strategy in 2019, which was the default social behavior was moving from town square, which was the old Facebook paradigm of, hey, we’re all in this together. This is our college network. This is our friend.

Ben: Hey network or public, here’s what I think.

David: Yup. And in many ways, Instagram was also town square.

Ben: Totally. The Instagram team noticed over time that even before Snap started eating their lunch, that engagement would decrease the longer you stayed an Instagram user, because you, over time, being done with these permanent posts and you saved it for the big announcements in your life, that there wasn’t a natural way to just effortlessly share, because when these platforms all started, everyone was in debt sharing with the town square the whole time and everyone was getting a little bit more clammy about that as time went on.

David: As the stage got bigger in the town square, it’s like well I’m not going to take the stage unless I really have a great performance that I’m going to give.

Ben: Every piece of data and metric that they had, you’re right, they realized, oh the world is shifting from town square to—

David: To living room is the way that Mark put it. That social was shifting from town square to living room.

Ben: By living room, he means small private groups of super close relationships.

David: And Snap and WhatsApp were both really interesting data points about this. WhatsApp obviously with private messaging. Snap also in its own way there was the original Snapchat use case of disappearing one-to-one photo messaging. But stories really is where it became clear, hey this is widespread and this is a big problem that is going to threaten Instagram, Facebook, the whole default motion of sharing beyond just a one-to-one messaging platform. I want that to be much more private and locked down too.

Ben: If you own an engagement platform and someone figures out a new mechanic to make them much more willing to freeform share, and your platform seems to encourage them to stay back, be quiet, only post once in a while, maybe lurk, it’s not good. Content creation on the platform going down is really, really bad if you are an engagement company.

Snap represents the idea that people are sharing way more if it’s ephemeral, and WhatsApp represented the idea that people are shifting the places they communicate from more public to more private, and from larger groups to smaller groups, both of which are concerning if you are Facebook blue app.

David: Now here’s an interesting question, though. How would Facebook know that attention is really shifting to these apps? Because remember, we’re now in the mobile paradigm, not on the open web where surfaces like comScore and Alexa and whatnot exist, and you can see traffic and engagement competitively on other apps. In the mobile ecosystem, it’s all siloed and locked down.

Ben: No, just Apple and Google really have that information.

David: But somehow Facebook was able to get pretty good insights on what was happening in the landscape.

Ben: Well, to have that you would either need to have some SDK that gets bundled into apps like an analytics provider, or you would need a VPN where the traffic was going through it so you could see the traffic.

David: Well, in 2013 Facebook acquired a small Israeli company called Onavo. Onavo was exactly what you just said, Ben, a VPN. Actually, I think it initially started as a data compression tool. If you were worried about your data plan, which was really important in many parts of the world, you could install Onavo on your device, it would compress the data, and then serve it to your apps.

Eventually that morphed into a VPN tool. The net of that is that they could actually see traffic and usage data going into specific apps on mobile phones in their network.

Ben: I think they looked at it as we need some way to level the playing field if we’re competing against Apple and Google in different ways, and they have this data because they own the platforms. We need to be able to see those same trends.

On the other hand, there is another way to view this. If you’re looking for examples where Facebook may have considered their own interests over being forthright with users over how their data is used, well this could be another big example. Users who were using Onavo didn’t download it with the intent of sharing their app usage data with Facebook.

David: They were just looking to compress their data or for a VPN service. Either way, Onavo ended up being a hugely important acquisition for Facebook because it really tipped them off to WhatsApp and Snap too.

Ben: Obviously, they fought those two companies in very different ways. One thing they learned from fighting Twitter over the years is that there are these social mechanics or perhaps an interaction paradigm might be the right way to talk about it. A post or a like or a retweet or a disappearing photo message.

The thing that matters is owning a valuable network. The idea that people are going to come and give you their attention and you own the place that they connect with other people that they authentically verifiably know. That’s the scarce commodity in order to win the engagement game.

David: And the mechanics are actually fungible.

Ben: And they’re totally a means to an end. if you discover some mechanic and you build this whole multi-hundred million user network ad platform based on it, that’s great. As long as no one comes up with a better mechanic than you and then goes and rebuilds the network somewhere else.

One thing they learned with Twitter was, hmm, they seem to be growing really fast with this status update thing. We need to look a lot more like status updates. And that worked pretty well. There were a lot of people that basically never switched to Twitter because they thought, oh I can just use Facebook for this.

David: Yup. I already have my network in here. I already have my friends. I already have my connections. Totally.

Ben: And it also has photos. Twitter’s this weird, esoteric text sharing thing. I’m not really all about that. I’m just going to keep using Facebook.

David: Oh Ben. Twitter did have photos for a while. It was called Instagram.

Ben: Well yeah, that API got turned off fast. As they’re looking around at Snap, hey, someone has discovered this new stories mechanic. My goodness, that is suddenly obvious that that is what the future is. It now feels old to do anything else.

It’s like when you got a Retina iPhone for the first time, and you’re just holding your non-Retina iPhone and you’re like, this is instantly a piece of crap. I’m not ever going to touch this disgusting thing again.

I think when someone invents a new interaction paradigm, it’s one of these things where you have to adopt it, because otherwise people are just going to flee. Of course, your business depends on you adopting it because you can’t let someone use this new discovered mechanic that’s perfectly timed for this moment in history, with these set of cultural acceptances, and this new set of technologies to go rebuild the network somewhere else.

So I think the thing that they discover is either through buying or through copying a mechanic, we need to protect our network by bringing these interactions into our family of apps either by, as Ben Thompson would put it, the audacity of copying well, or of course by buying them.

David: Yes. Copying well or buying well.

Ben: Listeners, of course, we don’t actually ever know what anybody’s intent is or what they’re thinking when they decide to buy a company or something like that. This is just David and I guessing at strategy from the outside. We do have, thanks to a court case, an actual email from Mark Zuckerberg on February 28th.

David: Oh yeah, this is so good.

Ben: 2012 to their then-CFO talking about the time they’re discussing the Instagram acquisition, but laying out the idea behind an acquisition strategy.

The basic plan would be to buy these companies and leave their products running while, over time, incorporating the social dynamics they’ve invented into our core products. One thing that may make neutralizing a potential competitor more reasonable here is that there are network effects around social products and a finite number of different social mechanics to invent. Once someone wins at a specific mechanic, it’s difficult for others to supplant them without doing something different.

It’s possible someone beats Instagram by building something that is better to the point where they get network migration. But this is harder as long as Instagram keeps running as a product.

Pause. Which is pretty interesting. That is the argument of why to keep Instagram separate and running as its own product. Instagram’s already discovered this fascinating new mechanic around publishing one image at a time with these beautiful filters. If anybody else tries to come after them, they’re already ahead. Actually, the best thing to do is own Instagram and let it keep doing its thing.

Anyway, resuming. Integrating their products with ours to improve the service is also a factor. But in reality, we already know these companies’ social mechanics, and we will integrate them over the next 12–24 months anyway. The integration plan involves building their mechanics into our products rather than directly integrating their products, if that makes sense, by a combination of these two things: neutralizing a potential competitor, integrating their products with ours to improve the service.

One way of looking at this is that what we’re really buying is time. Even if some new competitor springs up buying Instagram, Path, Foursquare, et cetera, now will give us a year or more to integrate their dynamics before anyone can get close to their scale again. Within that time, if we incorporate the social mechanics they are using, those new products won’t get much traction because we will already have their mechanics deployed at scale.

It is God damn brilliant, David. Here’s my commentary based on all this. It is so smart to basically say, well, if we buy them, we basically get two strategies that we get to execute at the same time.

One, we just leave them alone and let it keep succeeding. It’s actually not a current potential threat. We don’t know if it will turn into a potential threat. Hell, Instagram wasn’t making any money yet. WhatsApp didn’t have a feed. It was just a way people connected. It’s not like just because they had a network, they were going to turn it into a feed.

There’s this idea that it’s not a competitor now. By buying it, we basically get this option on if it becomes a competitor, if we completely leave them alone and let them decide what to do. But then there’s also, look, we will integrate those mechanics into our core product, Facebook, the blue app that already has well-built out network effects. And by owning this thing that could become a competitor, there isn’t white space in the middle.

Anyone just like us is also going to take time, they’re not going to do it as good as the original, and they don’t have our network. Therefore, whether the winner turns out to be the original product or us incorporating the mechanic into the blue app, we’ve won either way. probably what’s going to happen is both.

David: Yes, so brilliant.

Ben: And that is what happened with Instagram. WhatsApp is actually different. It never really turned into a competitor. It just serves a completely different use case and is also owned by Meta. Then with Snap, they tried to buy Snap over and over again and it didn’t work. So they basically figured out how to bring those mechanics into Facebook’s core products with stories in a way that Snap kept doing their thing. But there was really no reason to leave being an Instagram user because you already had that functionality with your own network anyway.

David: And here’s what’s really interesting about the “story” in Instagram, then ultimately coming to the blue app. Stories as initially launched by Snap in the Snapchat app was a separate tab with a list of your friends who had published stories. It was literally a table, like a list.

Ben: Oh, I don’t remember that really.

David: Yeah, yeah, yeah. No, it wasn’t. You know the paradigm we all think of today as the heads in the row at the top of your app, Instagram did that. when Instagram copied “stories” and launched stories, they launched it as the algorithmically-ordered row of heads at the top of your app.

Here’s where the internal advantages that Facebook, the family of apps had. They already had this world-class ranking algorithmic expertise within the company that was being applied to newsfeed, that was being applied to ads, that was already starting to incorporate early ML and early AI, start the beginning of FAIR (Facebook AI Research lab) that had already happened here. This was coming into the product.

Well now when you’re launching a new feature like this, in this case stories, you can say, oh great, is this going to be made better by our algorithmic and AI expertise? Of course it is. Let’s do that.

Ben: That’s interesting. Basically, even if you copy the feature wholesale, it’s actually even better not just because you have the network, but because your algorithm tech is very advanced. You can make sure it’s the most possible engaging version of that particular interaction design.

David: Yeah, so when Instagram, in those early days when it was launching stories, they were optimizing the ranking of heads at the top of the app by how likely you are to tap on that person’s head.

Ben: Of course they’re doing that. I do want to jump forward to TikTok.

David: Let’s do it because it’s a continuation of this story in a very different way.

Ben: In the mid-20-teens here, there’s a bunch of other stuff that we got to get to, but TikTok is particularly interesting because it blindsided Facebook. If you believe that the asset that matters is the authenticated real name network of people you actually know or people you care about following, you think, oh, I always have a lead as long as I can incorporate someone’s mechanic.

What if engagement is possible on an app that has nothing to do with your social network? That’s terrifying because this big asset that you’ve built, this durable competitive advantage of, we already have all the people so why would you want to go anywhere where your friends aren’t, the magic of TikTok was AI suddenly in the late-20-teens became sufficiently advanced that it could show you the optimal thing created by anyone in the world. Most of the time you don’t actually know that person.

Suddenly, the only competitive advantage that Facebook the company has is people have a habit of tapping on their apps when they’re bored. But that’s a pretty thin competitive advantage. That’s really easy to break.

David: Habits are changeable, especially on a mobile phone where another app is just an icon away.

Ben: Totally. So if someone is a great competitor like TikTok who is incredibly well-funded, very good at strategically buying ads, has created their own growth function that is they’re just a very different animal and they’re a very well executing machine. They obey a completely different set of laws, rules, regulation, norms being a Chinese company. That created a competitor for Facebook that is more significant than anything they’ve ever faced.

This wasn’t like, oh, there might be an existential threat from this in the future. This was like, oh crap. As soon as people form a habit around, oh, I just opened this black app with the white little music note on it, we have no more competitive advantages to throw at that problem. We must, as fast as possible, make something like reels just to stem the bleeding. And then we can figure out what to do from there and hopefully get to a market stalemate with them.

David: Which is basically what has happened. But interestingly, TikTok and ByteDance its parent company, as you point out, Ben, was so completely orthogonal to everything that Facebook had learned over the past 15 years because yes, number one, most importantly, it’s not social media. It’s media. There’s no social and your graph doesn’t matter anymore.

Ben: That’s a great point. It’s digital media, it’s user-created media, it’s mobile media, but it’s really not social media.

David: Yup. Doesn’t matter who your friends are on TikTok.

Ben: What matters is in the first 10 videos you look at, is their AI pretty good at picking up the stuff that you care about and the stuff that you don’t? It’s not can you find a whole bunch of other people you know?

David: I really want to keep hitting on this insight again, that Mark correctly identified of social shifting from the town square to the living room. This is a second order effect of that shift that the company didn’t see coming. Because once you shift social from the town square to the living room, it now becomes possible to divorce media from social. You’re already getting your social now in private, in your digital living room. The town square can become something that is completely not social.

Ben: It’s turned into this pretty interesting dual pronged approach that they now have, which is you open Instagram and you have reels. The job to be done by TikTok is now done by a thing that you already have a habit around. It’s bundled into the place where you actually consume your social media. Great. That’s step one.

Step two is, this is where it’s nice to have WhatsApp. This is where it’s nice to have Messenger for these living room style conversations with just a few of your friends. The new user behavior is, I see a thing that got millions of views. I can think of one or five people to share that with, and then I have a private conversation about this public object in private.

David: I will share that on WhatsApp or I will share that on iMessage or I will share that on, whatever my actual friend living room network of choice is.

Ben: And from where social networking started back in 2004 and ending up now in this reasonably complex interaction paradigm of I get AI served videos from people I don’t know, tailored to me, and then I privately share those in multiple groups with those who I love and care about.

David: That’s a huge evolution.

Ben: It’s crazy that we use the term social network or social media for these two things that are light years apart.

David: Totally. The social and the media is now completely divorced, and that is because of the shift from town square to living room.

Ben: Yeah. We really need to evolve our language around all this stuff.

David: Okay. That I think is dimension one that TikTok is completely orthogonal to Facebook. Dimension two is actually organizationally.

TikTok, as we chronicled in our episode about it a couple of years ago, came out of ByteDance. Well it was an acquisition of Musical.ly by ByteDance, and then ByteDance turned it into TikTok.

ByteDance had Toutiao before having TikTok. ByteDance is an AI company. ByteDance is not a product company. The core product that ByteDance makes is its AI along with all sorts of other stuff. Tiktok is the product vector, one of many and the biggest by which they deploy their actual product, which is AI.

This is very, very different from Facebook, which has always been a product company. Organizationally, the way ByteDance is set up is there are these centers of excellence within the company, well the algorithms, the AI that is controlled by a centralized team within the company. At least that’s my perception of how they’re different organizationally from how Facebook used to be.

Facebook and Facebook leadership was obviously watching this very, very closely. If you look at how Meta is organized now, it looks a hell of a lot more like that. Meta today is organized. There are centers of excellence across the company. Same way that I was just talking about ByteDance.

There’s FAIR. That’s the fundamental AI research team. There’s infrastructure shared across the whole company, deployed to all the products and apps. There’s revenue and monetization shared across the whole company, deployed to all the products and apps. There’s growth shared across the whole company, deployed across the various products and apps. There’s integrity and safety, which is a big, big, big advantage that Facebook has. They’ve reorganized how the whole company looks.

There are still all the individual products in the family of apps. Those all have their own heads and they all roll up to Chris Cox who has head a product for the whole company. AI too actually reports to him. But all the infrastructure, all the revenue, all the growth, all the integrity and safety are decoupled from those actual products now, and live in these centers of excellence across the company. And I think there’s a lot of ByteDance influence in that.

Ben: Makes sense. It’s the most formidable competitor they’ve ever faced, and they really have no way to neutralize it.

David: Yup.

Ben: All right. All this talk of TikTok, them being an AI-first company, clearly Meta today is an AI-obsessed company. How did that start?

David: Well, that is a great question, Ben. But first, we’ve got a brand new partner to announce that we are very excited to share with you all. Huntress. Huntress is one of the fastest-growing and most loved cybersecurity companies today. It’s purpose-built for small- to mid-size businesses, and provides enterprise-grade security with the technology, services, and expertise needed to protect them.

Ben: They provide a revolutionary approach to managed cybersecurity that isn’t only about tech. It is about real people providing real defense around the clock.

David: How does it work? Well, you probably already know this, but it has become pretty trivial for an entry-level hacker to buy access and data about compromised businesses. This means cyber criminal activity towards small and medium businesses is at an all-time high.

Ben: So listeners, Huntress created a full managed security platform for their customers to guard from these threats. This includes endpoint detection and response, identity threat detection and response, security awareness training, and a revolutionary security information and event management product that just launched.

Essentially, it is the full suite of great software that you need to secure your business, plus 24x7 monitoring by an elite team of human threat hunters, in a security operations center to stop attacks that software-only solutions can sometimes miss.

Huntress is democratizing cybersecurity by taking security techniques that were historically only available to large enterprises, and bringing them to businesses with as few as 10, 100, or 1000 employees at price points that just make sense for them.

David: In fact, there are over 125,000 businesses now using Huntress, and they rave about it from the hilltops. They were voted the industry leader in endpoint detection and response eight times in a row, and the industry leader in managed detection and response by customers in the G2 rankings for the second time this summer.

Ben: So if you want cutting-edge cybersecurity solutions backed by 24x7 expert humans who monitor, investigate, and respond to threats with unmatched precision, head on over to huntress.com/acquired or click the link in the show notes and just make sure to tell them that Ben and David sent you.

Okay, David. Meta today clearly an AI-obsessed company. How did that start?

David: I mentioned this organization FAIR a couple of times. I think it’s fair to go back and tell the actual origin of that because it’s really interesting.

Ben: It’s Fundamental AI Research.

David: Originally it was Facebook AI Research, and now it’s Fundamental AI Research, given the company is now Meta. Back in the 2013 timeline, when the company was finally fixing mobile coming out of the hellscape of 2012, of going through that transition during the IPO and the shift to mobile, Mark and the company’s VP of Engineering, Mike Schroepfer or Schrep as he’s known, were trying to figure out like, okay, we’ve finally got a little moment to breathe here. The tiger isn’t chasing us at the moment. Can we take a step back and try and figure out when is the tiger going to jump out next and try and get us?

Ben: Or all these crazy things are happening because we’ve been having to play defense as user attention shifts with this new paradigm of mobile. How do we not let that happen next time? How do we play a more active role in whatever the future of technology is, so we’re not getting whiplashed around, and we can control our own destiny a little more?

David: Basically, not actually owning the platform sucks. We’ve realized this now. How do we make sure this never happens again? That spawns some of the early work on AR and VR—come back to that in a sec. Even this early in 2013 on AI, and AlexNet had just happened right at this time. For a whole history of AlexNet, go check out our Nvidia series where we talk all about that and the importance to AI and ML. Facebook actually had a small ML team already at this point.

Ben: They were mostly doing computer vision. The early machine learning folks at Facebook were mostly looking at can we do automatic image tagging to reduce the, you remember how cumbersome it used to be when you’d upload a big batch of photos, then you had to go tag every single one, and then somehow magically—I don’t remember the exact year—it was all pre-suggested and you were like, whoa, this is pretty cool. That was what the original ML folks at Facebook were doing.

David: AlexNet and those big AI ML breakthroughs in that era were all around image recognition. Facebook was already attuned to this and like, oh, okay, there some pretty big technological stuff happening. Maybe this is going to be the next platform vector of the future. We should get serious about this AI stuff.

So summer 2013, Mark and Schrep go out to recruit Yann LeCun, and convince him to come start an AI research lab at Facebook. Yann was then and is now one of the top AI luminaries in the field, top academics out there. Yann says, well I’m flattered that you would want me to come work here. That’s interesting. But I live in New York, I teach at NYU, I don’t want to leave, I don’t want to quit teaching at NYU, and I’m pretty darn committed to open source and publishing all the work that we do.

Mark and Schrep are like, (1) sounds good. You can stay in New York. (2) Okay, you can stay at NYU too. (3) Yeah, we love open source. We were built on the LAMP stack. Great. Publish everything.

Ben: We haven’t talked about this at all at this very long episode so far, but Facebook had been doing a ton of open source publishing in basically everything they were doing. They were open sourcing things like Cassandra and their backend. They were open sourcing front end frameworks both on mobile and on web.

David: React. React Native, yup.

Ben: Exactly. They’ve been open sourcing a lot of the language modifications they were making. I don’t think open compute had started yet, but that was right around this time where they were starting to open source their data center designs.

David: Data center architecture, yup.

Ben: Facebook definitely has a very particular open source strategy that we’re going to talk about in analysis. But needless to say, they have been advocates of open source since the very beginning.

David: Yann is like, wow. Well okay, when you put it that way, sure. What’s really important about how they do this is this is not Microsoft research, this is not Blue Sky, yeah, let’s go focus on anything. This is a highly-focused AI research group. We are going to research AI.

I think that has made all the difference. It is super clear exactly how that is going to translate into helping Facebook and Facebook’s products.

Ben: This was pretty, I don’t want to say contrarian, that would’ve been wrong, but it’s one of a dozen specializations of computing that you could have glimpsed into for the future.

David: And it was one of those earned secrets that Facebook and Google had at this time—2013, 2014, 2015—of oh we can very, very profitably apply this work to our current products in social media feed recommenders.

Ben: It’s very funny that the rest of the world is having their AI moment now because Meta had theirs in 2013–2014 before language models, long before LLMs. Facebook realized they could profitably spend billions of dollars on AI systems to recommend: (a) what post you should see next in your feed, and (b) what ad out of the entire inventory of ads we could show you. What is the best one for you at this moment in time for that advertiser’s spend.

Today, there are a lot of companies who are spending on AI in hopes that the use case materializes in the future. Facebook’s is wildly proven and has been for a decade and is incredibly profitable for them. That’s just something to keep in mind whenever you’re looking at Facebook talking about AI today. Yeah, there’s a lot of future-looking stuff they’re talking about, but it is already super at scale, a great business, and has been for a long time. It’s not like they’ve been quiet about it, but no one cared until now.

David: It’s one of the fundamental advantages that Facebook has today relative to all other AI companies, is that they are guaranteed profitable ROI on all of their AI spend, all of their research, all their infrastructure, all their GPUs. They have highly, highly profitable ways to put it to work today.

Ben: It is a nice to-have for Meta if the next form factor of computing is an AI assistant like Meta AI. But it’s not essential. That’s one of many things that they could do profitably with AI technology.

The other thing you got to remember is 2012 Facebook, their engineering brand still wasn’t top-notch. Everyone knew that they were a up and coming great startup that went public, but they weren’t necessarily reputable in the computer science community as these people are Microsoft-level researchers to advance the state of the art.

David: Or Google-level in the way that Google always was from the beginning.

Ben: I think Mark and Schrep had to make this point to Yann, and then when Yann joined it made the point to the rest of the industry that this is not something that’s going to wax and wane with our financial results. We’re in for a decade on this particular area of research.

It’s amazing. It really only took two years to bear fruit for them before they had profitably deployed AI systems in production, because most of the time these research projects take much, much longer.

David: It’s funny actually thinking through this now. I think there are three dimensions to the payoffs from this investment in AI and FAIR back in 2013. One is the near-term and immediately an ongoing profitable paybacks to incorporating AI in existing Facebook products and social media feed recommenders. Highly, highly, highly profitable.

Ben: It just makes every ad that’s displayed more likely to make money, and it makes every feed post that is displayed more likely to get engagement.

David: 100%. Then there are two which are, hey AI might be the next computing platform. There are a lot of interesting stuff going on. This allows us to have a piece on the chess board and be a player in the next platform transition should it be to AI.

Ben: And this is just like what Meta did with threads with WhatsApp. With threads, they’ve said we’re not monetizing it right now. We’re going to see if it becomes a big close to a billion user platform. If it does, then great. We’ll figure out the right monetization strategy just like we did with WhatsApp. It may or may not be ads, specifically, but we build things that get engagement and then we later figure out how to make money on them.

David: Then three, they had no way of seeing this at the time, but FAIR and AI actually becomes the way that they can catch up and re-architect to compete with TikTok via reels. No AI, no reels.

Ben: That’s such a good point. If not for FAIR, TikTok could have disrupted Facebook.

David: Totally. Reels would not have come out as fast as it did.

Ben: Facebook would not have had that near-term weapon to go, huh. We need something that stops our users from leaking out and jumping over to this other app, for a use case that we basically can’t match. It is magic how the reels and TikTok algorithms work and now YouTube with shorts. Very quickly they do figure out how to put things in front of you that are incredibly engaging just for you. They call it the for you page.

David: And that have no social connection to you.

Ben: Exactly. You’re right. The decision looks more prescient than ever given that.

David: Totally. Okay, that’s FAIR in AI. Let’s jump forward now to 2016, 2017, 2018. I think for our purposes here today, it’s interesting to view the 2016 election, Cambridge Analytica, all the fallout that comes out of that which much, much, much ink has been spilled through this lens of the consumer social transition from town square to living room. Because I think it actually explains a lot.

Ben: Ooh. All right. I’m curious to hear what you’re thinking here. But to catch us up, here’s what happened. Until 2015-ish, Facebook could do no wrong other than all the data privacy issues that they had worked through with the FTC in 2011. But they are now increasingly finding themselves in very controversial situations that they are ill-prepared for as all of humanity is now joining the network. First of all, content moderation is becoming an issue.

David: Oh yeah, for sure.

Ben: Facebook is trying to figure out their role in this. Are they a neutral platform? Do they need to police what is said on their platform? On top of that, actually defining what is allowable speech on their platform is becoming trickier than ever, and it is hard to create one set of rules and abide by them globally. Facebook was just caught flatfooted in a big way by the fire hose of questionable content that people would post.

Ultimately they end up throwing huge amounts of people at this, hiring tens of thousands of contractors to deal with content moderation, building out really sophisticated policies, programs, escalation, and review processes. They have an ever escalating set of posts, videos, and live streams shared around the platform.

David: That’s right, because it’s not just text anymore. It’s obviously photos but videos now too.

Ben: Reviewing real time video, absolutely. As you would imagine, they are not perfect at this. They’re creating headlines left and right about objectionable things being shared on the network.

In building technology that enables everyone around the world to communicate with each other and organize, they also have the flip side of anyone can share anything and if it’s engaging enough, there’s a ton of eyeballs that see it. Misinformation is starting to become a big deal too. Ultimately, this comes to a head in the 2016–2018 period after they have whittled away a lot of the public’s goodwill. We’re just going to zoom in on this moment for the company.

The 2016 election happens and a lot of people are looking around for an answer to wait a minute.

David: Yeah, this isn’t the outcome that I expected. What happened?

Ben: Exactly. There are a lot of people who feel this couldn’t have happened on its own, legitimately. Somebody did something. One thing that absolutely did happen is there was a firm called Cambridge Analytica that sold their services to the Trump campaign, and had a methodology to create psychographic profiles that they believed would work that were derived from a Facebook quiz application. This ends up being an absolute honeypot of a story for anyone who is fired up about a big change to the country that they are not excited about.

There were basically four stories concurrently happening of how Facebook was involved in the election. (1) There was Russian interference. (2) People are spinning up fake news sites, not necessarily for politics but actually for the ad dollars generated on them. (3) Cambridge Analytica. (4) The Trump campaign may have been really good at using Facebook’s digital marketing tools.

David: I think that that is undeniably true.

Ben: There’s been years of investigation into each. It seems like that last one is actually the biggest lever. Yes, Russians did spin up fake pages and accounts with pretty modest spend, and people absolutely spun up fake news sites and generated fake viral stories for politics and also profit.

Cambridge Analytica, yes, it was a thing. But they ended up with actually pretty crappy derivative data from a quiz, not the treasure trove of raw Facebook data on you and your friends that everyone feared. But David, to your point, really what happened is the Trump 2016 campaign just actually took Meta platform seriously and got really, really good at using the tools.

David: Quite ironic in that Obama totally doubled down on Facebook as a strategy and used it to great effect to win the 2008 and 2012 elections.

Ben: As you can imagine, I’m preparing for this episode and I’m like, okay, I just heard Mark Zuckerberg say on another podcast. By the way, great interview with Alex Heath at The Verge after Meta Connect about this. He dropped this line, “People thought that all this data had been taken and it had been used in this campaign. It turned out the data wasn’t, and the data wasn’t even accessible to the developer.”

David, I’m listening to that and I stop because I’m thinking, well that’s not what anybody thinks because there was a ton of reporting on the fact that they did. So what actually happened here?

Here are the chain of events back in the 2010s. Facebook built the platform that we talked about, that was incredibly permissive in the early days when you offed your profile against an app, you gave it all of your profile data, and in those earliest days it could actually see your whole friend list too. Facebook eventually made their API much more closed off, but it took a couple of years to do that.

Now there was a terms of service that required app developers to delete any data they had after a period of time and only use it for very specific purposes. But of course Facebook couldn’t actually guarantee that people were complying with the terms of service. It was just, you were a breaching contract.

At some point a developer made an app that was a quiz, users could opt into taking that quiz. The quiz asked you questions and by combining the answers to those questions with the profile data you authenticated from Facebook, that app developer then tried to label you with a pretty basic psychographic profile. Then ultimately that derivative data, the labels of personality characteristics that were derived from Facebook data and your quiz answers, is what Cambridge Analytica had as best I could tell.

I read a good amount of the report that the UK government put out about this. The conclusion was that the quiz was taken by 320,000 people. Since those people had access to friends, the total surface area was the public profile data of 87 million Facebook users. They also found that Facebook had requested that Cambridge Analytica delete all the data, both the actual Facebook data and anything derived from the Facebook data back in 2015 before the election even started.

It was yes, true, an application got access to Facebook data on a lot of users. It’s also true that it violated terms of service to use that data for other purposes, and they did not comply with a delete request or they may not have complied. It’s difficult to reconstruct. On top of all this, it supposes that Cambridge Analytica method of taking these quiz answers and translating them into something that could impact voting, worked, like it was effective. That’s pretty thin.

David: That’s the most thin thing of all of it.

Ben: Here’s what the UK report concluded, and this is in government-speak, so it’s all very hedged. “While the models showed some success in correctly predicting attributes on individuals whose data was used in the training of the model, the real world accuracy of these predictions when used on individuals whose data had not been used in the generating of the models, was likely much lower.

Through our analysis of internal company communications the investigation identified, there was a degree of skepticism within Cambridge Analytica as to the accuracy or the reliability of the processing being undertaken. There appeared to be concern internally about the external messaging when set against the reality of their processing.”

At the end of the day, the Cambridge Analytica thing, in particular, was a nothing burger, but that is not the story that gets told. The UK regulator was given full access to this, and ended up being quite skeptical that the methodology even worked in the first place. They’ve got this crappy derivative data, not complying with terms of service requests to delete it, and the method may not actually work.

Now this doesn’t exonerate Facebook in any way from what could have happened. They had an incredible treasure trove of data that apps could slurp up in the early 2010s. But the actual narrative of Facebook data that Cambridge Analytica had impacted election results is wildly oversold.

David: All that aside, as we were wading through all of this, trying to make sense of it, and think about how we’re going to talk about this when Acquired, there is such an enormous, enormous disconnect between reading what actually happened and then how people felt and what the lived reality of the brand impact on Facebook was. That was huge. Nothing could have been huger.

Ben: To me, the story here is there sure was a lot of ill will and discontent with the Facebook brand that people were willing to dive into head first on. It’s almost like Facebook isn’t guilty of whatever the heck people think happened with Cambridge Analytica all these years later.

David: But something is very, very wrong that hundreds of millions of people believe that.

Ben: And it’s not just wrong with the system. They definitely, in behaviors along the way, managed to earn people’s distrust. I think that’s the issue.

David: So let’s bring it back to my supposition here, which is that I actually think this paradigm shift that Mark identified from town square to living room was way more right and way more powerful than even he realized. This is one of the second order effects of it. The quiz that was the app on Facebook. That happened way back in the early platform days. Totally solidly in the town square days.

Then the election happened in 2016 when things are already starting to transition. WhatsApp exists, Snap exists, Snap has launched stories, Instagram has now copied Snap and launch stories in that very same year.

Two more years go by the Cambridge Analytica news doesn’t actually hit until 2018. By 2018 we are solidly in the living room era. All of the supposed real or imagined impacts of this Cambridge Analytica thing had their roots back in the town square era. We’re now here in the living room era, and that is what people’s expectations are of Facebook, of social media, of everything that they are sharing in their lives.

I think this is to my mind at least, where the disconnect is. People’s expectations shifted. But now all of a sudden all this stuff from the past becomes really, really relevant to the present. When you look back at the past through this lens of what the present is, you’re like well this is really effed up. Like how the hell did the company do this?

Ben: The answer was they were trying desperately to grow. They were trying to will a platform into existence. Of course the cultural norms were a little bit more permissive, but the cultural norms were really never permissive to the point to say, sure you should download and store a big JSON dump of my whole profile information, any pages that I like, and anybody whose names I’ve connected with. That always was probably a little bit of an overstep, but then once developers were able to do that and then violate terms of service and store it forever, then it just becomes this ticking time bomb that’s out there.

David: Exactly. It’s like this weird temporal and cultural like mismatch of it all landing in 2018 when the world is a very, very different place.

Ben: In part, you can see why Facebook massively prefers the ad-based business model of we never even expose any information to you advertiser. You target and then we just deliver the ads. You never get to know anything about who the users are on our platform. We don’t sell data.

That is a true fact whenever they stand up there and say we don’t sell data. They don’t. They sell the opportunity to target users. But in fact if they did sell data, it would be a bad thing because someone that could then build a competitor. The asset that they have is actually their data that they choose not to sell for business reasons.

David: Totally. Then there’s this whole other layer too here of, well why did they do this in the first place? Why did developers have access to so much data? They were building platform, they were building an entirely different business with a different strategy. Now they’re an advertising- and media-based business. We’re almost talking about two totally different companies here as I think we’ve told in the story.

Ben: I do think, when I search my heart of hearts though, when I was developing apps in the 2010 period and would let people use Facebook Connect or whatever the current branding of accessing Facebook accounts was at the time, I was always like, whoa, this is a lot of data that comes with someone clicking that button.

David: Well I think you made the point of Facebook had to sell the platform to developers.

Ben: Totally did.

David: They didn’t have the underlying hardware, technology, and operating system to offer, so there had to be a really big sweetener.

Ben: I just can’t believe it, though, after all these years and all this writing and all these headlines. If Cambridge Analytica were more competent, they probably could have had more impact. But as it stands, I just can’t believe there’s not a big story on the front page of major newspapers with big headlines saying, actually Cambridge Analytica didn’t really have that much in the first place, their methodology appears to have not worked, and mea culpa, sorry for getting you all all riled up. If you don’t want to like Facebook, sure. But go pick a different reason.

Okay, so what happened with all this? It is worth knowing to finish the story earlier from the consent decree in 2011 that said Facebook is going to be carefully monitored for privacy concerns. As you can imagine, the Cambridge Analytica news coming out put a gigantic bullseye on Facebook.

The FTC said, aha, you guys signed a consent decree a mere seven years ago. What the hell? This is exactly what we’re talking about. Facebook just decides, you know what? What’s the big dollar sign, that cheque that we can write here? Because there it seems like we just want this to be done, we want this to be buried. We are a different company now. We want to put that to bed.

The whole settlement here was not specifically about Cambridge Analytica. There were a bunch of other things that are privacy-related, but the number is $5 billion. David, there is one thing that it comes with, and that is a 20-year window that they are monitoring Facebook. That is exactly matching to the words that Mark gave us on stage that Cambridge Analytica, the way they handled it was a 20-year mistake.

David: Now, I think by the 20-year mistake, he meant much more of the damage to Facebook’s brand that came out of this than specifically the FTC settlement.

Ben: It’s funny. I was thinking this is a very Facebook-appropriate comment. If we had subtitles for this episode, this one would be, it’s complicated.

David: I guess. So great.

Ben: It really is.

David: So great. It really is. Really is.

Ben: What it comes down to is I think there’s an ethos at Facebook, that at many steps along the way, there was an opportunity to grow, to win, to compete. What mattered was winning. Then when the dust settles, you can look around and say, okay, what was the impact of that? That is one way to run a company. The other way is to be really careful and ask permission.

Facebook fell into the first camp and they probably wouldn’t be here today or certainly at least wouldn’t be a global player with four billion users today if they hadn’t fallen into the first camp. But by falling into that first camp, you have stuff that comes up where you’re like, ah, crap, I wish we hadn’t done that, and this is one of them,

David: The time bombs. Totally.

Ben: Okay, so what actually happens? Facebook becomes a lot less valuable after all of this comes out. The company announces in their July 2018 earnings call that they’re going to be more focused on user privacy, and that revenue may slow as they make this intense focus. They dropped 19% in one day. They wiped out $119 billion in market value which was the largest single day loss for any company in history at that time.

David: At that time.

Ben: Save at that time.

David: Yeah. To be exceeded in the future.

Ben: So this whole privacy issue, the $5 billion settlement, all this stuff that just happens with Meta, it has big ripples for the whole tech industry. There’s a phrase that is known on the lips of every American, somewhere in the neighborhood of, Cambridge Analytica, election interference, fake news. I hate social media, Facebook did this, YouTube did that, blah-blah. This is a persistent drumbeat that is underneath the whole national discourse.

There is an opportunity to be the anti-Facebook here if you want to be, to be the company that says, we are so unbelievably, incredibly about privacy even more than we ever have been before. David, who does that?

David: That is Tim Cook’s Apple.

Ben: Yes. Apple, a company that I love and adore and was actually wearing an Apple T-shirt earlier today, and had to take it off before recording because I felt like it was weird to be wearing my Apple T-shirt when recording the Facebook episode. Apple is a company that has always been incredibly privacy-conscious. They both use that as something they believe deeply in their soul and design into their products, and is amazing for all of us who use their products to get that privacy and have that trust.

They also, to use Ben Thompson’s parlance, use it as a strategy credit. There are areas in which it really behooves them because they don’t need to do server side stuff, they don’t need to do advertising because they make a lot of advertising indirectly from Google, and they don’t need to do a lot of other things. They can tout, hey, we’re unbelievably secure with your data. We take privacy incredibly seriously. No one takes it more seriously than us.

They start really beating this drum and by 2021 they decide, you know what? In iOS 14.6, we are going to launch a new policy called App Tracking Transparency.

David: Yup. ATT.

Ben: What that means is, a few years ago we mandated that anybody who is tracking you across apps start using something called the IDFA, the identifier for advertisers.

Now in the past, Apple had let you just actually reach in and grab the device’s unique identifier, which was pretty cool because as a developer, it was unrelated to advertising. Think of it almost like a serial number of this device. You could use that for things like, hey is this the same user across multiple applications? I’ve got my SDK and my code running in multiple developers applications.

I can do interesting things like say, hey, this person both takes, runs with Strava, and they also use Google Maps, and you can just gather data across apps. Similar to how Facebook was gathering across the web with like buttons everywhere or with Facebook Connect everywhere, so they could build this holistic profile of things you do off of Facebook.

Apple stops letting you use the device identifier. They force you into using this IDFA. Then with iOS 14.6 in 2021 they say, hey, if you’re using IDFA, part of that API is that now whenever someone launches an app for the first time, it’s going to ask them in this really aggressive language, are you okay with getting tracked or do you want to ask this app not to track? What do you think people are going to click?

David: There’s a default answer to that question.

Ben: What actually ends up happening? Most people click ask app not to track. A whole lot of Facebook’s targeting basically falls apart. They no longer have a picture of you outside of apps that they actually own. A lot of the reason why advertisers can get so good at targeting is because of this holistic picture that is built for you across your phone.

David, this is the example manifest in practice of the thing you were talking about all the way at the launch of mobile of, what’s one of the reasons why Facebook’s beautifully constructed business model doesn’t work in the mobile walled garden ecosystem. It’s because the operating system maker can make a change like this that just affects core functionality that you are relying on, and now you don’t have access to that data so you can’t run as effective of an advertising service.

David: A minute ago when you asked the question of who is banging the privacy drum and I said it’s Tim Cook’s Apple. One of the reasons I said that is that, actually I didn’t know this till doing the research, Mark and Steve Jobs had a great relationship, which makes this story all that much more interesting.

Ben: Mark is very Steve-like.

David: Very, very Steve-like. They actually met for the first time in 2007, right as all of this is coming together, and the first meeting was that Steve actually reached out to Mark and wanted to meet him and build a relationship, which I don’t think Steve did that with too many people.

Then on the back of that, the next year in 2008 when the iPhone SDK was launching and people were building apps, Mark and Facebook actually go to Apple and meet with Steve and they pitch Steve on exactly what you were just talking about, Ben, like hey, how about you let us bring our platform that we’ve just built that has all these developers into iOS into your platform.

Ben: Really? Oh, to make Facebook platform part of the iOS developer. Whoa.

David: Yes. Which from Facebook’s part is like a Hail Mary. They’re probably not going to say yes here, but this is really big risk to us. We might as well try.

Ben: But they did have enough of a relationship where Facebook and Twitter both were privileged citizens on an early version of the iPhone OS. Like n the setting screen, even before you installed any apps, there was a Facebook and a Twitter settings for (I guess) native integration between the operating system and those networks.

David: And social networks, yes, totally. Steve, supposedly the legend goes, Mark and the Facebook team make this whole pitch, and at the end of it Steve just looks at Mark and is almost a little sheepish and it’s just like, thank you. We’re just not going to let somebody else build a platform on top of our platform.

Then he pushes back in his chair and it’s clear that’s the end. Then he is like, hey so by the way, what do you guys think about Microsoft and working with them? Then they just have a really nice conversation supposedly for the next couple of hours, shoot the shit, and talk about Microsoft and computing history. It’s this very cordial relationship. I think the sense I get is that Steve really liked Mark and respected him.

Ben: They’re both pretty product visionaries. They’re both very stubborn about their views of the future, and they both were very right about their views of the future. I can see it.

David: Totally. And clearly that is not the feeling today between the two companies. I think what happened since, okay, if it had just ended at Apple isn’t going to let Facebook build a platform on top of their platform, that might’ve been it. But then when mobile app installs happened and Facebook all of a sudden…

Ben: Facebook was making billions of dollars off of deciding what apps in the app store get downloaded. Apple had to have felt, hey this is actually…

David: This is our turf.

Ben: Ours, yeah.

David: Then after everything that happened in 2016–2018 Cambridge Analytica privacy, to your point Ben, part of this is that Apple has always cared about privacy, it’s always worked well for them, and it’s part of their main product strategy. Part of it, too, is oh great. Here is the ultimate competitive vector that we can have against. Somebody we really don’t like, who we feel has profited off of our platform unfairly.

Ben: You end up with Apple feeling like everything that happens on our platforms is ours and these are our users to protect, and no one’s going to do wrong by our users in any way. You have Apple who is protective as all hell, and then you have Facebook where Mark Zuckerberg more than anything in the world, wants as much freedom to operate as possible. You have Apple trying to constrain, and you have Mark who hates feeling held captive.

David: Now here we are in 2021 and Apple’s like…

Ben: We’re making our move.

David: Facebook’s been reeling. We’ve been beating the privacy drum for the last couple of years since 2018 and Cambridge Analytica. Now it’s time to make the move.

Ben: What ends up happening, actually the first couple of quarters, not much. Facebook’s talking about our earnings calls, hey we think this is going to have impact. It’s not huge. But then February 2022, which is technically the end of year 2021 earnings call for Meta, they drop the bomb.

Interestingly in question and answer—I listen to the whole earnings call—the CFO casually says in a response to someone, oh we think that the impact from ATT to our ad business will end up costing us on the order of $10 billion for 2022. What? That is eye-popping. What ends up happening is there’s a 26% drawdown in a single day.

David: Remember we said to put a pin in the single day drawdown records?

Ben: Yes. The market cap went from $900 billion to $700 billion. The actual number is they lost $232 billion in market cap. The new largest in history surpassing their previous record. Also on this call, they announced their first ever quarter over quarter user decline.

Keep in mind what’s happening is on top of Meta basically saturating most of the Internet-connected world by this point, TikTok is also really, really peaking. You’ve had some marginal users using the app less because they’re moving to TikTok, and Facebook is reacting to TikTok and trying to put reels in. They’re cannibalizing their own revenue by encouraging people to watch these short-form AI-recommended videos, that actually don’t yet monetize as well as the newsfeed.

There are three fronts that are destroying them here. There’s ATT, there’s TikTok competition, and they’re making revenue-cannibalizing changes in their own app. The hit continues. It traded all the way down 46% by April 27th, and then ultimately it bottomed on Halloween that year—this is eight months later—with a 72% drawdown.

David: Ooh.

Ben: They lost 72% of their value between February and Halloween.

David: Ooh. All in the year 2022.

Ben: It’s crazy.

David: And that is like Nvidia levels of drawdown, back in the crazy Nvidia journey that we talked about.

Ben: It’s totally insane. The interesting thing is at this time the real threat wasn’t actually ATT. The biggest of those three threats was TikTok stealing users. The whole ATT thing is about how much money can we make off of an ad because it is so well-targeted. That’s an optimization. That is useless if you do not have users to advertise to in the first place. So the actual real existential threat is TikTok.

Ben Thompson makes this really great point, and I know I’ve quoted Ben over and over this episode, but I think he’s just been so astute on Facebook at many points throughout history. “This decision to make these product changes to respond to the TikTok threat in the face of ATT to do these at the same time is a founder-led decision. If you had a professional CEO, the correct thing to do to preserve your job and shareholder value is to wait 6–12 months before you start reacting to TikTok to let the whole ATT thing blow over.”

Mark’s like, I can’t get fired. I think the right thing to do is react to TikTok now because every day this problem compounds and gets absolutely worse. I don’t care that there’s this horrible narrative going on right now with ATT that is going to cause us apparently to lose $10 billion of revenue we otherwise would’ve gotten this year. We must cannibalize revenue in addition to that to make these product changes.

What it ends up with is a 72% drawdown, and what it also ends up with is the chance for it to 5X from there, which is what has happened. The company saved itself by acting correctly in this and they had to go through this wild tumultuous two-year journey in the process.

David: It’s just another episode of Mark making these sets of decisions because he’s Mark, because he controls the company, and because he has the stomach to withstand it.

Ben: What ended up actually happening from ATT, it’s been value destructive overall because I don’t think the amount of money that has shifted away from Facebook has been captured by Apple. I get the sense that the app install business for the app store, those search ads is going great, but it’s not equivalent to what the monetization over at Meta is on app install ads or was on app install ads.

David: I think Apple was hoping in its heart of hearts that this would shift a lot of those dollars into ads in the app store, and the reality is people just don’t go to the app store searching for apps anymore.

Ben: Not to mention this kneecap Meta on a lot of ads that are not app install ads. We have a friend that’s in the retail business who was saying that their ads doubled in price in 2022 when this happened. They were scratching their head, and they have nothing to do with the app store.

All that happened is that entrepreneur said, I guess I’m going to keep advertising on Facebook as a platform. I hope it gets better. I’ll pick some other platforms, but I can’t advertise with Apple. There’s nothing for me to do. It’s a retail good, it’s a physical item, so I guess I’ll just pay more money to acquire customers now.

It ends up actually hurting the business when go-to-market channels get less efficient unless you can shift that spend to a place where your customer is also hanging out.

Ben: In practice, Meta is fine. When a shift like this happens, the most scale player with the most engineers and the most data turns out they’re still the winner. Meta launched this thing called Advantage Plus, and now if you’re an advertiser they use a whole bunch of other data signals.

You still have a customer acquisition budget, you’re going to spend it, you’re probably going to go spend it on Meta the same where you were before, and it’s just not quite as efficient as it was. Oh and by the way, anyone else that got hurt from the changes with ATT, just doesn’t have the ability to react and build around it the way that Meta had because they don’t have the resources Meta does.

David: All of this finally brings us to the last piece of the story here that I’m sure listeners at this point you’re probably saying, what the heck? We’re however many hours in, you guys haven’t said the words Reality Labs or Oculus or AR or VR once.

Ben: Oh, is Facebook in that industry?

David: Yeah. You know? I guess they are.

Ben: All right, so what happened here? Way back in (call it) 2012, Facebook starts getting interested around the same time that they’re starting FAIR for AI, in what is the platform of the future. They start doing some hardware prototyping on their own. They’re not really set up for that yet, but they do want to do the same approach that they did with AI which is focused research, not general research. Pick a particular thing where we have an opinion about something we think is going to be the future, and then invest deeply in it.

So a crazy thing happened in February of 2014. They opened up the purse strings and paid $19 billion for WhatsApp. They had so much conviction that the future was AR and VR, that 34 days later they paid another $2 billion for Oculus. I had no recollection that these things were a month apart.

David: No I didn’t. I knew they were both in 2014 but I didn’t realize they were a month apart.

Ben: It’s wild. I think Mark got the demo of Oculus and was just like, oh this has the most credible potential to be the future out of anything that I’ve ever tried, so yes.

David: But I think actually the right way to think about this is through the same lens as FAIR. It’s a research lab in the same way that FAIR is a research lab. And it was like, oh great, $2 billion to kickstart the research lab. Fantastic.

Ben: All right. You and I have both tried Orion. It is unbelievable. Fantastic. Very clearly a path to the next generation of computing device. I don’t know if this is going to be the winning company, I don’t know if that’s going to be the exact winning device, but never have I been so sure that a mainframe very far away from me, to converting to a PC which lives three feet from me, converting to a phone that lives in arm’s length from me. The next logical step is glasses that live on my face.

I used to sound ridiculous, then you and I tried Orion, and now I’m like oh yup, that is going to replace or augment my phone in whenever these things are commercially available and consumer grade. I’m glad we waited do this episode because I think I would’ve had a pretty different take having only tried VR headsets, Vision Pro, and big googly things over the years, I think I would’ve been unconvinced frankly because I didn’t think it was possible to put something in that small of a form factor.

That is the current product experience that you and I have recently had. Now let’s look at the business strategy and the financials of how we have gotten here and why. There are two ways to look at reality labs. The first way is to answer the question, what would have to be true about the business to be great on its own and justify all this investment?

David: And let’s level-set on how much investment has gone into it.

Ben: Yes. Based on their spend already since they started reporting Reality Labs as a separate segment in 2019, they’ve spent right around $60 billion. That is in operating losses for the segment.

David: And does that include the $2 billion for Oculus? Or either way, it’s a rounding error.

Ben: Who cares? I don’t think it does. With that level of investment, it already needs to be essentially the most successful and profitable consumer product in history to pay itself back. That is the only possible outcome here where we even get our money back. I know this sounds wild but that is the bet. No other outcomes are acceptable. Fortunately, if we want to model this out, we know the financials of product like that. The most successful and profitable consumer product in history, the iPhone.

As a thought experiment, what if Meta managed to launch such a product, say the Orion glasses, tomorrow and say that such a product grew at the exact same rate and with the exact same profit stream as the iPhone? Well I did that napkin math, and if 2024, this year was Apple’s 2007, you just take all the iPhone’s cash flows and you start the clock right now, Meta’s cumulative cash flows from Reality Labs would be net negative until at least 2035.

David: Okay. We’re talking 11 years.

Ben: You would get back to break even on your investment 11 years from now if starting tomorrow they managed to create the most profitable and widely adopted product in human history.

David: Which obviously is not happening tomorrow.

Ben: That also assumes generously that Meta could build a services business attached to it the size of Apple’s services business, which probably generates right around an equal amount of profit. Basically, take all the profits from iPhone and double it. That’s actually what you would need. It’s actually fair to attach a similar size services business.

David: I think that’s very fair, knowing Meta and all of its capabilities and everything that we just talked about.

Ben: That is the bet. Let’s just be super clear. Anything else is a complete incineration of cash.

David: That is the bet looking at through the lens that you posited of a purely financial investment perspective.

Ben: Yes. The second way is actually financial, too. I think there’s probably a third more emotional way. The second way though is if you’re Mark and you constantly live under the thumb of platform control, you’d do almost anything to get out of it. That’s not irrational.

Apple made a $10 billion dent in their revenue just two years ago with ATT. They could—I don’t think they would, but they could—at any given moment just pull you out of the app store and you’d have little recourse. That is an existential business risk. It’s an unlikely one, but every day you could wake up and all of your access to all Apple customers could be over.

David: Which, by the way, we haven’t talked about it, but Google could do the same thing.

Ben: Google could totally do the same thing. They have the right to distribute or not distribute anyone’s app in their store at any given time. Look at Apple and Epic with Fortnite.

David: Totally.

Ben: There is actually an expected value calculation you can run, which is my entire company’s market cap times the likelihood that it could happen, which it’s extremely low likelihood, but because the market cap is $1.5 trillion, the expected value is still a very big number.

Then if you’re thinking this way, what is a reasonable percent of your market cap to invest every year in a hedge that might might get you out from under the thumb of big tech platforms? At Meta being worth $1.5 trillion If they’re spending $15–$20 billion a year on Reality Labs losses, that’s a little over 1%. Is that worth it? If you truly believe that this is the most effective way to offset your most existential risk in the next two decades? Hell yeah it is.

David: Hell yeah it is.

Ben: Now, am I certain that this is the best way to hedge that risk? No, I don’t know. Is there a huge amount of execution risk along the way?

David: Well they’re running many hedges on this. Of course, this is part of what FAIR is, this is what Llama is, et cetera, et cetera.

Ben: But it’s a 1% tax. It’s generous to frame it off of market cap. You probably should frame it off of revenue. But still a 1% tax off of your entire enterprise value every year? Nah.

David: I think there is also potentially a third way to frame it. It is funny, I think you put it in the emotional lens. I actually think less emotional but more upside-oriented of Facebook did have the taste of being a platform for those couple of years there from 2007 to (call it) 2010–2011.

If they own the next platform, not only will they do everything that you’re just talking about of like, oh they’ll have an iPhone-sized business because if it turns in to be glasses and Reality Labs, they’re going to be selling the hardware and maybe they can make that profitable like the iPhone is profitable, et cetera.

Then there’s also your point number two, which is this is a defensive play. Our core business right now is at risk because it’s built on somebody else’s platform, and Mark made the comment on stage with us that they’ve run the analysis. They think they could be twice as profitable in their core media products if they were on their own platform. Sure, great.

But then, there’s the big, big play is the third part of, well what if we actually built a platform again where we have millions of developers who are developing on us, and we are participating in their business in the way that we used to participate when they were developing on us in the past? That’s an even bigger play.

Ben: And to underscore your point, David, just a few weeks before we recorded this episode, Meta held a multi-hour keynote at Connect in September 2024. Of those multiple hours, zero minutes were dedicated to their core products of social media apps, their $100 billion business of selling advertisements on these social media products. At least with developers, it’s not about any of that at all. Literally the entire keynote.

David: Yeah it did. It was great.

Ben: It’s Meta AI, it’s the Llama models underneath it. It’s the open source strategy, it’s developers building for the Quest. It’s announcing new products like the next iteration of the Ray-Ban Meta glasses. It’s revealing Orion. Zero minutes to their current products or business. And they’re one of the biggest businesses in the world.

David: Yeah, wild, right?

Ben: They are all about this platform future. Here we are trying to make the case of, they did it. They decided they were an ads business. They built one of the two greatest advertising systems ever known to man, and one of the most amazing business models ever. At their annual developer keynote, they’re talking about how they’re going to one day become a platform company.

David: There we have it. That is the story of Meta. I should probably say for a minute here, too. My experience using Orion—you were talking about yours a minute ago—I thought it was unbelievably compelling. Before trying it, I never would’ve believed that any type of AR or VR would really have been the platform of the future. Then after trying Orion I was like, oh yeah, I could wear these on my face all day and I could replace a lot of what I do on the phone with just these lenses being in front of me all the time. It would be way more efficient and way more enjoyable.

Ben: I would never have guessed that in this year you could fit that into glasses. Glasses and a wireless puck.

David: Totally.

Ben: Listeners, we’ve never shared our impressions on this. It is wildly compelling, and if you asked Mark, he probably would frame this whole thing differently than we have. It’s not about this hedge. He would say some things about platform control, but I think for him is just this general belief that we want to make awesome products, I believe there’s an awesome product to be made here, and I’m going to assemble the best people I can to go work on it.

I think that’s the most interesting duality of this company, is it’s both. It’s what is the best strategic move to make, to marshal my resources, and a much more touchy feely, like I want to make products that are great and bring people closer together because that’s the mission of the company.

David: Well put. Should we wrap the story there?

Ben: We should. All right, so we will catch you up on the business today just to put some numbers to all of this, and then we will move into analysis. As of the end of reporting last quarter, there are 3.3 billion daily active people across the whole family of apps.

Okay. Astonishing. That’s up 7% year over year. The family of apps revenue per person is about $12. The end of year stats from last year when you just look at the Facebook app, not the whole family of apps, the daily active users are 2.1 billion. of those 3.3 billion daily active people across all the apps, 2.1 billion are on Facebook, and the monthly number of Facebook is 3.1 billion monthly active users. So 2.1 daily, 3.1 monthly.

WhatsApp and Instagram are in the neighborhood of two billion monthly active users. WhatsApp has 100 million now in the US, which is a big narrative violation that WhatsApp will never catch on in the US, and iMessage is dominant, and even after that it’s text message. This is crazy. A hundred million people a month in the US use WhatsApp now. It’s been a slow burn over time growing and growing and growing, but to me that came out of nowhere.

The other narrative violation here, there actually is a lot of growth among young adults using the Facebook app itself in the US. I think a lot of people think that’s sort of a boomer thing.

David: It’s funny. One thing that may be contributing to that, which we didn’t put in the story, is marketplace. For me personally, I have totally boomeranged all the way back around on the Facebook blue app and I am now a loyal DAU because of marketplace. It’s awesome.

Ben: Which is crazy because you don’t use any other form of social media besides posting on Twitter for Acquired, right?

David: Correct.

Ben: They’ve announced that Meta AI is on track to be the most used AI assistant by the end of the year. It’s worth disambiguating Meta AI from Llama. Llama is the name of their family of open source models. Those models do power Meta AI, but Meta AI itself is a branded consumer experience that lives both standalone and in a bunch of Meta’s apps.

Their revenue in 2023 at the end of the year was $135 billion, operating income of $47 billion. That’s a 35% operating margin. Worth knowing, just like all tech companies, they have become CapEx-heavy the last few years. Last year, they spent $28 billion in CapEx, which you should mostly read as data centers.

They operate a hyperscaler-sized data center footprint, give or take. You think AWS, Azure, Microsoft, and Google Cloud, Meta is the fourth one. They just don’t sell it to anyone else. It’s only consumed by internal teams, so huge amounts of investment in AI hardware and just other data center expansion.

The balance sheet is fortress. They have $50 billion in cash and $58 billion, including cash equivalents and marketable securities. They have 71,000 employees and their market cap is $1.5 trillion up from $230 billion in just October of 2022.

David: Wow. Incredible.

Ben: Totally incredible. Two really insane observations about the state of the company today. All of their products seem to increase user engagement over time. All of these products have different use cases, and it happens across geographies. There is something in the water at this company.

Why is it that WhatsApp and Instagram are both increasing in user engagement over time? It’s the growth function. It’s this purpose-built heat-seeking missile of attention and metrics that the company pays attention to, where across a whole broad product suite, engagement increases. It’s not like they have one thing that happens to do really well. It’s a process.

David: I would say it’s a whole set of things. The company calls them centers of excellence internally. It’s the growth function, it’s AI, it’s the ads, and the revenue team. All of this contributes to it.

Ben: Yup, fair. The second insane thing, at IPO, the US and Canada market had an average revenue per user of $11. That number is now $227.

David: Wow.

Ben: Yeah. And globally, that average revenue per user when you include all the emerging markets and less valuable markets for them is $44. They really, really monetize now. How much of the world does Meta really have left? To put a bookend on this.

David: I almost asked you this question all the way in the beginning when you were teeing up the episode, but I’m glad we’ll save it for here.

Ben: According to the UN, last year there were 5.4 billion people online. This includes China. So they’re basically saying two-thirds of humans are online.

David: And Meta does not include China.

Ben: Yes. Meta has four billion monthly active people across the family of apps. China alone is 1.4 billion. While this isn’t totally exact, I think you just apply the same multiple to China and say, well two-thirds of humans are online, two-thirds of China is online. That means that there are 940 million people online in China.

David: That’s probably being conservative. I think you could probably argue a much higher percentage of China is online.

Ben: But let’s be conservative. That leaves about 450 million people or 6% of the human population who have access to the Internet but are not yet Meta monthly actives. That doesn’t mean they’re not Meta users. That just means they weren’t monthly active users as of the end of the last reporting period. Meta’s addressable users who aren’t yet users is less than 6% of humans.

David: Just wild. And then still, I think the craziest thing about the company is that if you look at the daily active user figure, it’s not that much lower.

Ben: Right, and across all products. So once you frame it this way and you’re like, huh, there’s only 6% of the population left either through reactivation or signing up that they could get, you understand why they put so much effort behind emerging markets, behind doing things like zero rating, doing custom deals with telcos, rolling out fiber, bringing countries online for the first time even when they have no near-term monetization potential. internet.org was the name of their initiative for a long time around this to basically say, look, we are saturating humans. We got to figure out how to get more humans on the Internet.

David: Yup. Move on to analysis?

Ben: Yup. Okay, in the analysis the first thing we’re going to do is power and then playbook. We are going to do a seven powers analysis of what enables Meta to achieve persistent differential returns, or to put it another way to be more profitable than their closest competitor and do so sustainably.

This is interesting. Who is their closest competitor? I think that’s probably worth defining first and foremost. Ultimately, they are in the business of selling advertising. I think their closest competitor is Google.

David: They are certainly the most similar looking businesses in their current forms.

Ben: But who do they compete against for the same profits or the same potential profits?

David: I’m not sure that it’s actually Google. It almost certainly is YouTube within Google, but I actually think it is other social platforms.

Ben: You’re right. I suppose it’s other places people spend time.

David: Yeah, it’s where people spend their time, online and offline.

Ben: I think really what has to happen here is an analysis of each set of stakeholders individually, like you almost want to do a seven powers analysis on the user side also, of why would someone pick a Meta product over a Snap product? And even though they’re not voting with dollars, it’s almost like their attention is a proxy for dollars because you just assume that those companies should do a comparable job monetizing the attention. But what makes sense I think is to just walk through each of them.

Counter positioning. They probably don’t have much counter positioning in the current state of Meta, and hold on AI for the moment. In their startup days, they did a lot of counter positioning against other global social networks by being a college-only authenticated social network. They were accepting lower growth, they were accepting a capped ceiling of number of users, and they were doing that because they wanted to make the trade that they felt a closed community is more important.

David: With higher engagement.

Ben: Scale economies?

David: Scale economies 100%. Absolutely.

Ben: This company’s in the business of scale economies.

David: Yes. I think that scale economies on the infrastructure side, on the GPU investment side. I think that scale economies on the advertiser side and on the ad experience side. Just by having so much more scale, they are the default standard way to spend money for advertisers.

Ben: Think about on the tooling side, think about the experience of being an engineer at Meta, and the thousand engineer years of work that comes out of that company every year on making the developer experience of working there better. It’s crazy with the revenue scale that they have, how much they can amortize these fixed costs.

David: Totally.

Ben: Everything at this company is scale economies.

David: Yes. Okay, scale economies for sure.

Ben: Switching costs. As a user, you’re pretty locked in once you have followers.

David: Oh, I hadn’t thought about that. Yeah, that really speaks to again, another shift we didn’t really talk about in the evolution of what we call social media is the creator aspect of all of this. The whole idea of a creator and influencer didn’t exist in the early days of social media and Facebook, but now you and me, our platform on the podcasting platform, we have incredible switching costs. Somebody who has a ton of Instagram followers has incredible switching costs.

Ben: Yup. Now fortunately, in the land of creators, you don’t actually need to switch. It’s just an and. It’s not like you’re ripping out one vendor and putting another vendor in. I do have a sunk cost in building a following on a given platform, but that doesn’t actually prevent me from also launching on another platform if you have the time to do that.

I think there might actually be process power. Normally there’s not, but I keep going back to this, like there’s something in the water. Their products grow in engagement over time. Their growth team does 10,000 little things to open up every step of the funnel as wide as it can be, and make the most frictionless fluid experience for users.

David: Interesting. I was actually having this discussion with Hamilton recently about the difference between process power and cornered resource. A good test for this is when you bring somebody brand new into the organization and plug them in, does the power transfer to them as well? If no, then it’s a cornered resource. If yes, then it’s process power.

In this case I think that’s potentially true. Yeah, you can bring in new engineers, you can bring in new product people, and transfer the advantage to them.

Ben: For a while they had legitimate process power in how they shipped. Statsig has started on the premise of this. It blew my mind when someone from Facebook came and gave a talk at Microsoft in 2014, on how they ship product and how it rolls out.

Not only the feature flagging elements of it, but they can watch performance auto roll things back if they’re causing negative performance metrics. How there’s just deployed code all the time with a whole bunch of experiments turned off. How there are these experiments that are running in different ways in different markets, and they can statistically significantly disentangle which results are from which experiment when they have multiple experiments that are concurrently running with the same user base. That was real voodoo that only Facebook did for feels like a decade.

David: I suppose Google probably does similar type things in different ways but nobody else really does.

Ben: I think OpenAI does. I think OpenAI has so much former Facebook DNA and thinks the same way that I think they do a lot of this thing too.

David: Okay. I can buy that there’s some process power here.

Ben: Branding. Branding is such an interesting one. Can they have negative brand power?

David: Anti-power branding weakness?

Ben: There’s no love for the brand Meta, there’s no love for the brand Facebook, there is love for the brand Instagram.

David: Yeah, but diminishing as it gets larger. Certainly not like it once was.

Ben: There’s reliability with it all. There’s branding that comes with being any big company that you’re large, trusted, and institutional.

David: But I think this is a clear example of how strong the other powers in this company are. That despite the brand taking a 20-year hit in Mark’s parlance, they’re still today a $1.5 trillion company.

Ben: Then cornered resource, unless you’re going to call Mark one, which always feels a little bit too cute to name the founder.

David: I actually think there are some cornered resources here within Meta. Ironically, I think one of the biggest ones is the integrity and safety team, and all of the privacy controls and work and security that they have done.

Ben: That they’ve had to harden over the years, yeah.

David: Even though this is also the source of so many criticisms about the company…

Ben: It’s now pretty robust.

David: If you think about anybody that would start up to try and compete with them—somebody I was talking to put this to me—imagine a startup trying to build a hate speech classifier in Farsi. Meta’s got that.

Ben: And trying to have the relationships with the public policy people in 200 countries to understand what is acceptable speech in each of those countries.

David: Yup. While probably YouTube to a certain extent, TikTok perhaps, also have similar strengths here, I doubt there is any other company in the world capable of actually operating privacy, integrity, and security. And within that is anti-spam stuff too, which is super, super, super huge at the level that Facebook is.

Ben: It’s interesting. Okay, so in looking at this, why does Meta do $47 billion a year in operating income and why do we all believe that’s going to continue for a while? What actually is the defensibility here?

Is it the network economies? Because they’re so tautological. I would argue it’s not that because what we’ve seen is TikTok showed us there’s a way without initial strong network effects to go capture people’s attention, and thus eventually the ad dollars.

David: Yup, and they competed very successfully with Meta in doing so.

Ben: If this were five years ago, I would’ve been like, why even talk about the rest, network economies? Once you have the network, then you know the rest doesn’t matter. That’s just not true anymore.

David: Well I think the network economies are still true relative to anybody who would compete in a social product.

Ben: Yes.

David: Which still exists.

Ben: Totally.

David: Social is just now divorced from media.

Ben: And most people do use both, and sometimes they even use different apps for it. But it’s two different modes. It may be the same session on the couch, but you do need both use cases of, show me who the people that I intentionally follow, what they’re doing, and then show me entertaining things.

David: Totally.

Ben: So part of their business is defensible from the network economies. But the other part, honestly, it feels like habit.

David: Well habit, but I think scale economies are really big for that too, especially in AI. I think the only way that this new paradigm exists is because of AI and because of GPUs specifically. I’m not even talking about generative AI. I’m talking about feed recommender systems and personalization systems. There is a very, very large minimum barrier to entry there that keeps going higher as ByteDance and Meta keep investing in it.

Ben: That’s a great point. Okay. Playbook?

David: Playbook. Let’s do it.

Ben: Let’s talk about what Meta is doing with AI and that will lead us into our first playbook theme. We’ve talked about so far the beginning of FAIR, them starting all this AI research, the early 2014 on use of feed recommenders, and the AI for the ad matching system. But there’s a lot going on with Facebook and AI right now that we really haven’t talked about.

There are two words for you to know. One is Llama, and this is the family of foundational models that Meta has developed and these are competitive with OpenAI, Anthropics, Claude and…

David: Google Gemini, et cetera.

Ben: And then there’s Meta AI. Meta AI is a consumer brand that is the way that you interact with Meta’s self-hosted version of Llama, or maybe give it a little bit more credence than that it is an application that Meta has, that uses llama in the background, but provides Meta-specific AI experiences, some of them bundled into apps like in WhatsApp chat or in Instagram…

David: The blue app or whatever, yup.

Ben: Yeah, but also there’s a Meta AI website that you can go to and interact with it directly. But Llama is the models themselves, Meta AI is the consumer product. Llama, they have spent billions and billions of dollars, huge amounts of R&D, huge GPU clusters to train big data centers. Interestingly, it is all open source.

David: Mark makes a big deal about this.

Ben: It’s interesting to think about why. It’s not open source in the same way really that, oh Linux is open source and this is the free standard that everybody uses and there’s just a foundation behind it all. Meta is putting huge amounts of CapEx and OpEx, huge amount of dollars into willing Llama into existence. It’s very different from these cheap, grassroots, open source projects of the past.

David: It’s open source like Android is open source.

Ben: Yes. So why are they doing this? Well, if you ask Mark why they have an AI model at all, and I’m quoting from a great blog post that he put out about this, “We must ensure that we always have access to the best technology. and that we’re not locking into a competitor’s closed ecosystem where they restrict what we build.”

Okay, I understand that. You believe that the most important thing is to control the key technologies that make your products possible. Okay, that last bit is important. Basically what Mark is saying here is we’re going to spend a lot of money training these foundational models, but unlike all of the competitors in the AI space, we actually don’t have a business model around making money on this. We’re going to spend all the money, we’re going to give it away free. Why does that make sense?

David: It’s Open Compute Project all over again.

Ben: Yes. For anybody who’s not familiar, Facebook made this move in the early 2010s, where they realized they were spending tons and tons of money on their data center infrastructure. And the vendors who they were paying—these integration partners, the server companies, and the networking companies—were making fat margins. They were thinking, this is dumb. We’re a really big customer, why is everybody else making so much money on us?

They looked around and they saw all these other big data center companies. They’re like, geez, all those people are paying big margins, too. What if we just publish the specs for the billions of dollars of work that we have done to make our data centers, then we start this thing called the Open Compute Project, and we just get a bunch of other people to adopt it too?

Well suddenly then, the Open Compute Project is this standard by which all the hardware manufacturers and integrators actually have to snap to because all the customers are saying this is what we want, and it’s a pretty genius way to drive margins down for these suppliers.

David: And also a way to guard against somebody in your supply chain building platform power.

Ben: What they basically learned from this is, oh we open source this thing. A lot of our costs went down because a whole ecosystem started using the thing that we open sourced. Even though we’re not making money—they’re not a cloud company, they’re not selling access to their data centers to anyone—

David: They’re not a hyperscaler.

Ben: It pays back in the form of saving them money. It can do the same thing in AI. They publish a really expensive open source foundational model that is in the conversation to be as good as these other closed source ones.

Well now, there are a lot of developers out there who are just going to build on the open source free one. It becomes something that the community can build on and improve and make better.

As Mark said in our conversation with us on stage, there are a lot more smart people outside your company than inside it. But effectively, what it does is it puts pricing pressure on the AI model companies to the extent that Mark views it as a super important key ingredient to the product experiences he wants to build in the future.

It’s super bad if there are a few closed source providers who can provide that experience and they: (a) lock him into controlling, so here’s what you can build, here’s what you can’t build, but (b) take his margin. Basically say it’s really expensive every time you want to make a call to one of our proprietary services.

When you think about it, it’s actually a form of operating leverage, where he is basically saying there’s a big fixed cost I am willing to bear in order to bootstrap this ecosystem and commoditize all of these compliments, commoditize all of these other closed source AI models. In exchange, what I’m going to get for that is just pricing pressure on all of them, so that in the future my variable costs are lower. I just get to keep more of the dollars that we bring in, rather than having to pay them out to proprietary model providers in the future. It’s a pretty novel business strategy.

David: Totally.

Ben: So I said commoditize your compliments. This notion was dreamed up and named by Joel Spolsky in 2002. There’s a great blog post about it. He makes this analogy, think about cars and gasoline. These are compliments. When sales increase in one sales increases in the other. You have a car, you need gasoline.

Well AI models end up being a compliment to Meta’s products. In order for them to build better AI or even the feed recommendation stuff we talked about, they need best in class AI models. And just because cars increase the sales of gasoline, that doesn’t tell you about how profitable an automaker gets to be versus a gasoline maker.

Imagine the automaker decided that they want to get into the gas business so that their customers could have access to low-cost gasoline. Or even further, let’s say the automaker decided to make low-cost gasoline just to drive all the other gasoline prices down. Well if gas is cheaper and you’re the automaker, you can actually charge more for cars. Since consumer willingness to pay is around the total cost of ownership, not about a car or gasoline specifically.

David: As a consumer you’re always buying a solution. You’re buying transportation or you’re buying a technology product doing a job for you. And if AI is a required input to that solution, then yeah, you’re thinking about that as a total purchase decision.

Ben: Or an advertiser is not going to pay you more or less depending on how much you have to pay the AI model provider. They’re going to pay you an amount, and if you want to maximize the amount of that you get to keep, it behooves you for you to not have to pay as much money to AI model providers.

David: Better put.

Ben: How does this apply to Playbook? Well, Meta seems to like taking all the risk in situations like this, putting lots of dollars in so they can take more of the reward. You could imagine a more moderate company saying, oh well there are going to be lots of AI vendors out there. We could just let them take care of that as their core competency, and we’ll just buy off the shelf from them.

But that’s not how Meta works, especially with everything they’ve been through with Apple. They have gotten true religion and I think always wanted to be in this position and they just have the capital to do it now. They will control the key technologies that matter to them both for getting to own and dictate product roadmaps and products decisions, but also for the financial upside of making sure that they control their own destiny and no one in the ecosystem has extreme leverage over them.

David: Only a company of the scale of Meta can run this “playbook” because no startup is going to be like, oh, okay, what is my strategy to make sure that OpenAI and Anthropic don’t get leverage over me? Good luck with that.

Ben: Totally. My first big playbook theme is Meta discovers commoditize your compliment, and is now looking for ways to use it everywhere.

David: I love it. Well put. Okay, great. I think my first one is almost a different way of putting that, which is really like Mark and specifically his superpower of placing multiple bets on multiple chess boards, never wanting to be backed into a corner. I think he has done incredibly well and I think the story of Meta shows all throughout, even in the very, very beginning as a tiny project, not even a startup where degrees of freedom were limited. He was always making choices and playing the game such that there were multiple options of how things could go. Right,

Ben: Mark is a master at maximizing his degrees of freedom and setting up the board such that in an uncertain future, there are multiple paths to victory no matter how the world unfolds. Which is unbelievably Gates-eyin. I feel like we said the exact same line about Bill Gates in our Microsoft episodes. The real comp for this company is Microsoft.

David: I think that’s right.

Ben: That was actually my next one. They do iterative product development. They put the first version out just to get feedback, see how they need to rev it, and get better for the future. The early days are characterized by hiring all the smartest people and prioritizing IQ over everything else. Commoditizing your compliment. Obsession with building a platform that other developers build on top of this whole thing about multiple bets in an uncertain future.

It’s funny. They were building Messenger internally, they bought WhatsApp. I bet they’re glad that they had that dual prong strategy. They were developing an app called Photos at the same time that they bought Instagram. That’s what they do. It’s Gates-eyin.

The most interesting thing that I think hit me in the face like a ton of bricks, Zuck after all the 2016 election fallout, the shift in public perception, and having to do all this testifying, is like watching an alternate future for Microsoft where Bill Gates had decided to stay at the helm instead of leaving after the DOJ case.

David: That is the biggest difference, is they both went through this hellish period. Mark came out of it and said, I’m more bought in than ever, and Bill came out of it and said, Steve, you’re the CEO now.

Ben: That’s exactly it.

David: Are you going to talk about Mark’s amazing line from the Harvard CS 50 lecture in 2005?

Ben: I’m not, but lay it on me.

David: Oh, it’s so great. Mark’s what? 20 years old at this point, maybe. He spent the summer out in Palo Alto. He gets invited back to do a guest lecture as a practitioner in the field at Harvard CS 50 class. He’s talking about product strategy. He talks about how he really admires Microsoft’s product strategy, and he thinks it makes a lot of sense of the first version is getting something out there and it’s usually not very good, but by the third or fourth version it’s pretty darn good. That makes a lot of sense to me as a good product strategy.

Ben: It is amazing. He lays it out right there. We don’t have to speculate on it. It is explicitly stated.

In this same vein, it is painfully obvious when you look at this company. That companies are just founders extended. The culture of this company is just Mark, and it’s a huge lever for Mark to act. This has been true over and over again, Microsoft, Nvidia, and Nike. It’s just how great companies are in the world. It’s very hard for me to point to a truly great company where the DNA isn’t like a lever on the founder’s personality.

David: I think that is right. I think it is doubly so with Meta because of Sean Parker and because it was set up from the very, very beginning that nobody could ever get rid of Mark, that it was always going to be his decision.

Ben: And there’s got to be a lot of companies out there that we’re looking at the success case.

David: This works if the founder is really, really, really smart and right almost all of the time.

Ben: As we talk to people, here are the traits that we heard over and over again. Mark is a genius, a really good listener, a fast learner. He goes from knowing zero to mastery in months or years. He has low ego about being right. That’s not to say that he has a low ego, but he has a low ego about being right.

He is obsessed with finding truth and open to being wrong. He’s intensely competitive, he’s relentless, he’s actually a very good product designer, and understanding the computer architecture that will be required to accomplish such a product experience at scale.

All these things if you have these characteristics, and you’re then empowered to singularly control the company at massive scale, yeah, it’s going to go well. But the first thing is actually much harder than the second thing.

David: It’s not a causal relationship here that having founder control will lead to the right decisions.

Ben: This is my regular reminder that in studying these episodes with extreme survivorship bias, we are looking at the most extreme outliers who, in every dimension you can multiply by, they’re at the edge of the distribution. Mark is, I don’t know, six or seven standard deviations from the mean human across the important traits that mattered to making Facebook. Oh, and by the way, with the right timing and the right luck and the right circumstance and the right know-how, my biggest lesson from doing an Acquired is these things are unrecreatable.

David: Agree. All right, what’s the next one?

Ben: Durable executive team. It’s shocking how many of the people that are Mark’s direct reports and their directs have just been there for a really long time. This team knows how to work together.

David: Oh, yes. I was actually talking about this with one of the execs at Meta the other day and he brought this up and was saying that like, this really feels like something very, very unique to Meta. I thought about it for a minute and I was like, actually no. I think this is something that is unique to all great companies. If you look across the landscape out there…

Ben: Apple’s certainly like this.

David: If you look at Nvidia, if you look at Apple, if you look at Microsoft in recent years, if you look at the original Microsoft…

Ben: Costco.

David: If you look at Costco. This is almost always true, really is true in every great company that I can think of. The core senior most executive team are all people who came up in the company or have been there for a very, very long time. It’s not a whole bunch of mercenaries there at the top.

Ben: That’s interesting. Okay, that brings me to this next one, which is something we talked a little bit about with Mark on stage. I’m curious to hear your answer to this, now that we’ve had all this time and space to think about it.

Here are all the battles where Meta has either won by buying the company, one by beating the company, one by copying or gotten to some sustainable stalemate. Myspace and Friendster, Google+, Twitter, Instagram, WhatsApp, we didn’t talk about this one, but Meerkat and Periscope?

David: Oh yeah.

Ben: They launched Facebook Live when that was supposed to be the next big thing. Snapchat and TikTok. That’s like seven, eight. Why do they keep winning? I have some thoughts.

David: It’s funny. As we were preparing to interview Mark at Chase Center last month, this was the question that we kept asking people. In some ways, I think it’s a little bit of a Rorschach test. Everybody’s answer was different, and it was always what they think is the most important thing.

I think my answer for this is going to be a version of Mark as the answer. It goes back to Adam D’Angelo’s Friendster testimonial about him being way too lucky and then realizing it’s not luck at all. I think it is that, as Adam was telling to me as he reflected on this, it’s a very specific way of carrying yourself through the world where you don’t hold too tightly to the path you’re on.

Obviously, Mark is incredibly driven to the goal of connecting all the people in the world and doing that through Meta, but he’s very, very open and flexible to the exact path that it’s going to take to get there. That combination I think has led over and over and over again to these threats emerging and Mark and the whole team there at Meta saying, okay, what is the way that we can neutralize this threat or defeat it? And we are open to anything on that front.

Ben: And I think they are open to being whatever they need to be to make that happen. I think Meta might look extremely different 20 years from now than it is today, almost unrecognizably because this company moves like water in response to whatever the new shape of the world is.

David: Hell, after trying Orion, I think there’s a very plausible reality five years from now where this company looks very, very different than it does today.

Ben: It really is this idea Meta is a technology company through and through, and then they leverage that technology to be whatever the hell they need to be to adapt to the new world.

David: Which is so different from all of the other big technology companies out there. They are all wedded to a particular vision of what they are.

Ben: Totally, because these are all different strengths. They aren’t afraid of copying. Okay, that is a strength. But that’s also a completely different thing than I’m going to place bets on what I think is going to be the big technology wave of the future and spend tens of billions of dollars on that. That’s a different way of winning.

Or in some cases I’m going to try to leverage my existing network effect to make sure I adopt someone else’s social mechanic. Okay, that worked in a handful of these scenarios, but that’s not at all what works in others.

David: They don’t just have one singular playbook that they run over and over again.

Ben: Right. This company moves like water, and product is an act of discovery there. I’m convinced it’s not pure invention.

David: You asked Mark that on stage

Ben: And there are some stuff they have to invent, like Orion you have to invent. But software where you can ship and quickly respond to user feedback and iterate, it’s like they’re chiseling away at the marble to find David. I don’t think they have David in mind when they’re starting most of the time.

There are a few moments where they did. Newsfeed is completely one of them. Inventing the social feed is a completely distinct and brand new thing that Meta created. It’s actually one of the few.

When I was a little bit more bearish on the future of Reality Labs, I was trying to come up with, has Meta ever successfully created something new that has become a profit center for them that is not adopted from someone else?

David: Well, I think the tagging of photos was pretty novel.

Ben: Yes, and newsfeed.

David: Yup.

Ben: But it’s funny that we’re naming those things as…

David: Platform as short-lived as it was.

Ben: No. That was a we wish we had an operating system, but we don’t, so let’s see if we can convince people that this is a sufficient platform.

David: Well, I think building a real honest-to-God platform with hundreds of thousands if not millions of developers making money on it on the open web.

Ben: Okay, fair pushback.

David: It was totally novel. Nobody had done it without an operating system before.

Ben: Maybe there was going to be a social platform in this new era. Maybe that was an operating system of sorts, or on par with operating systems to be able to create a platform on top of.

David: But I take your point. It’s certainly not every turn of the game where they’re inventing. They’re doing a lot of discovery here.

Ben: And maybe all companies do. There’s not something distinct to Meta here and everybody learns from each other, and that’s fine too. But that was something that I was racking my head on thinking through.

In fact, there’s even a testimony—this is funny—in the Mark Zuckerberg versus the Winklevoss case from way, way, way back when. Mark even makes some comment about, well actually the idea for Facebook wasn’t even new because MySpace and Friendster existed. It’s like this interesting positioning of the whole thing itself is actually a borrowed idea, which of course served him well in that particular case.

But the point stands. So much of this is borrowed and we just live in a world, especially in social, where you do have to just, maybe this is media as a whole, observe what the new format is and adopt it as quickly as possible.

So speaking of trying things, I just wanted to take a moment to honor all the failures. I kept a running list as I was doing my research on products that either died or just didn’t live up to the hype. Facebook Live, Facebook Watch, the solar-powered drone that they developed to provide internet access.

David: Oh yeah. That’s right.

Ben: That was a whole big thing for a while.

David: Yeah, that’s right.

Ben: Mark was actually two for three on, he was forecasting in 2015, what are we going to be focused on 10 years from now? That was one of them. But the other two were augmented reality and AI. It was an impressive call.

Portal TV, Workplace, which as far as I can tell, Facebook is the only company in the world that actually uses Workplace. No other enterprise has ever adopted it. I’m sure that’s not exactly true, but that feels approximately true. Beacon Facebook Deals, which was their Groupon clone.

David: Oh, that’s right.

Ben: Think about how nimble Facebook is that they’re like, ooh uh-oh, this social deals thing seems to be gaining traction. Maybe that’s a core part of the platform that needs to be a piece of this in the future. They just had zillions of these.

Facebook Gifts. Do you remember gifting? You could use Facebook Credits to pay for them. Facebook Credits. On top of Facebook credits, Libra?

David: Oh boy. Yeah. Wow. We didn’t even talk about that whole chapter on crypto. Yeah.

Ben: They invented a whole new cryptocurrency, a big consortium around it, a huge set of investors, and other Fortune 500s. Originally Facebook Messages was going to be a Gmail killer. You could email people into Facebook Messages and use that as an email suite.

Facebook Places. Foursquare was almost on par for a period of time with Instagram and Twitter as plausibly the next social mechanic checking in places. Facebook places was a real effort. Hey, now it just turns into tagging location on posts. But a first class post in newsfeed for a while was a check-in on Facebook Places.

And then there are all the independent apps—Lasso, Poke, Slingshot, Photos, Hello, Facebook Gaming, Life Stage, Moments, Notify, Facebook Watch, Moves—all the things they acquired—TBH, Beluga, they launched IGTV. This company tries everything. They move like water to discover what they need to be through an ever-changing environment.

David: So true.

Ben: All right. That’s it on the failures. What else you got?

David: Well, almost the opposite of the failures. It’s quite ironic that Facebook became the prototypical startup, given that the goal was never to be a startup. The goal was first to be a project and then to build…

Ben: A very profitable business.

David: The largest empire in the world, yes. There was no in-between of we are building a startup here. It was like, college project, oh, let’s connect all the people in the world and let’s get big really, really, really fast and skip this whole small company thing.

Ben: That’s such a good point.

David: And all this whole generation, multiple generations now of startup founders that have really embraced and romanticized this whole startup thing. On the one hand, it’s great. Silicon Valley’s so much bigger. There’s so much more investment, there are so many more VCs and all that. But it’s also just all ironic that it’s glorifying this startup phase when that is not the point.

Ben: Deeply. Okay. then we’ve touched on this one a bunch, but I want to put a pin in it. Why has Facebook always been in a precarious position? Why do they need to keep fighting these existential battles? Why are they so obsessed with building a platform?

I think the answer is they want to be as durable as a company that makes hardware with an operating system, that all the users use and all the developers have to target because all the users are there. I think that’s what they really want, but because they’ve never had quite that much defensibility, like, oh, a network effect is good, but it’s not as good as that incredible platform durability. They’re always trying to expand and be more every time they bump up against someone else, it creates a problem for them.

I’m using a bunch of weird metaphors here, but it just seems like the place that they occupy in the technology stack is just not quite privileged enough to do the things that they want to do, so they’re always at the whim of someone else knocking them around.

David: I think that’s right, and it will be very interesting to revisit this again in 5 or 10 years.

Ben: Yup. Part II in 10 years.

David: Yeah. Right.

Ben: And then as we drift to a close here, the engineering culture and being a technology company at their core has been essential. Early on, they really did manage to hire only A+ and then stay A+ after that forever and ever and ever. It was just this badge of honor if you were an engineer at Facebook. For product design too. They just had such a great talent density.

The set of things that they did on the technical side were over and over again a way to have their cake and eat it too. If you can move faster, you can learn more through your multiple iterations. Speed of development comes from having great tools.

One very great shining example of this is something called HipHop for PHP. I know this is very esoteric, but in the late 2000s they had this crazy idea that what we should do, instead of switching to C or C++ or Java, we want our engineers to keep writing PHP.

They wrote a compiler to C, so they didn’t have to take the performance hit from running PHP, which was an interpreted language. But it also scald this scalability problem because then they didn’t need to go hire all these systems-level programmers. They could hire web developers who wanted to move with that pace and flexibility while also having the infrastructure to run these massive systems and scale really efficiently.

That solved the scaling and performance problem. But again, in 2014 they realized, oh crap, we’re big enough. We now have enough sensitive data that we really should switch to a statically-typed language like Java or C#. But again, they didn’t want to force their engineers to learn that. People who code in those languages have a different culture than existed at Facebook, too. They couldn’t really recruit people. So they invented a new language called Hack that was very similar to PHP, but had static typing.

Tau is another example, this NoSQL database. They just keep finding ways where they invent new technology to solve a problem that probably only exists for them. Then they create this whole boutique system that allows: (a) having world class talent, (b) to have a ton of them, (c) everyone gets to move fast, but then (d) it’s all unbelievably performant and efficient, and they just don’t have to make trade-offs. It is wild.

Okay. Then my second technical one in addition to this is they are their own customer. We’ve talked about this. They’re a AWS scale technology company, but they don’t take outside customers, so they only have to build for their own internal use cases.

Now this sounds great, but it actually does have these big trade-offs. You can’t dramatically change what your infrastructure is used for. It is purpose-built, but it does let you be incredibly efficient and have high performance if you have good communication between the customer—the app team or the backend service—and the designers of that data center.

This is completely the opposite of Amazon. Amazon uses an interface so teams don’t have to talk to each other. At Meta, they require incredibly tight communication. It’s a very different organizational philosophy where they’re like, no, no, no. Not only do we not have external customers, we want this insane type coupling between our infrastructure and our internal customers.

Anyway, my last one, this is a company that grows intentionally. It would be easy to look at this company and say, wow, what a viral product. What a universally applicable product. That is not the case. It is unnatural to have connected four billion humans. This is a freak of nature. This is not just something that people adopted.

While I think it is totally fair to say, wow, it just blew up at Harvard, Facebook is the story of 50 different growth tactics in different eras, all carefully constructed and iterated upon.

Wait lists at colleges, strategically launching at every single college, picking each one for a reason, translation, internationalization, zero rating with carriers to bring new people online, figuring out when they needed to acquire versus build, carefully split testing every change, aligning the whole company on specific networks, on specific metrics.

Building relationships with governments in all these different countries, and at the very least, complying with local laws on where should we have certain speech laws versus not. it is completely unnatural for them to have done what they’ve done, and it’s all been very, very intentional to connect the world.

Last playbook theme I’ve got, there’s always another battle for Meta.

David: It’s funny. I think you could be listening and think like, oh yes, there’s always another platform battle ahead for Meta and that’s probably true, but I think you actually mean another societal battle ahead for Meta, right?

Ben: Totally. Now that they’re through the user privacy issues and the many years of the whole 2016 election conversation we talked about, and—I’m not going to list them all here—the many, many societal conflicts that they’ve had over and over again, the current issue for them is around the impact of social media on mental health, and in particular teen mental health.

David: Totally.

Ben: It feels fitting to put this near the conclusion of the episode because while going deep on this wasn’t a part of our understanding of how and why Meta as a business works so well, it is a really important topic. There are a lot of people making arguments that social media is bad for our brains, and the consequences if that ends up being globally true, is catastrophic for Meta. Probably a bigger challenge than they’ve ever faced at any other point in history.

If you’re asking yourself what are the things to keep an eye on going forward for them, it is of course all the product innovation, the growth of the existing business, trying to invent the next platform and everything we’ve talked about.

David: Orion and everything.

Ben: But it’s also how the mental health issue, understanding all that unfolds, and how they handle it.

David: Totally.

Ben: All right. Time to land the plane?

David: Let’s land the plane.

Ben: How are you liking this, by the way? This land the plane way of finishing episodes?

David: Well I actually have a proposal for you Ben. I don’t think we have landed—pun unintended—on the right nomenclature here or the right construct. We’ve been calling this land the plane, takeaway, the splinter in our minds.

Ben: What’s the one thing that’s knocking around for you?

David: What is the essence of this company? So I propose that we change the name of this final segment to Quintessence. What is the Quintessence? Yes, this is my inner French literature major coming to bear here on Acquired listeners. What is the essence of this company? We’ve just spent all this time, all these months studying this company. What is the very essence of this company that makes it different from any other out there?

Ben: All right. I like it. The quintessence. Ultimately, my big takeaway is the company moves like water. It is the company that has connected the world, that will always gear up for the next battle and be whatever they need to be in the next era. Whether it’s them defining the next generation of computing or creating all these AI experiences or fending off the next TikTok or the current TikTok, they just move like water.

David: I totally agree. I can’t think of a better characterization of this company and how they got to be so darn important in the world. Perfect place to end it.

Ben: And ultimately it is still very much a Mark Zuckerberg production.

David: Yes, that too.

Ben: All right. Quick carve outs?

David: Quick carve outs. My quick carve out is a Google product, actually, NotebookLM. This is freaking wild. Our friend, Ben Cohen over at the Wall Street Journal who wrote the great piece on Acquired a few months back, texted us, what was this, two weeks ago maybe? A week ago. It was like, have you guys tried this thing NotebookLM. I think you had, and me being me, I hadn’t. I was like, no, let me check it out.

I uploaded just the links to the sources that I used for my side of the research for the Microsoft Part I episode. Just links. And what it spat back out at me, I was like, holy crap.

Ben: At this point, I think we’re well past the Turing test, but it is the most convinced that something is actually a person I’ve ever been. If I didn’t upload all the sources and know that it was unbelievably tailored content to the thing that I just uploaded, I’m not sure I would know that it was AI. It’s pretty amazing.

I have two. One is a documentary on Netflix called Mr. McMahon. I was not a pro wrestling person growing up. I want to go be a pro wrestling person now. This documentary is incredible. It is some of the best storytelling I’ve ever seen.

Interestingly, it’s a documentary that is told with no narrator. There is story arc all throughout the episode exclusively with interview answers. You almost don’t notice, at some point you finish an episode and you’re like, wait. There was no narration in that. There was no cheesiness. It was all first-party accounts and then cuts to old footage of things that aired on TV.

Then the credits come up and of course it’s a ringer production. Bill Simmons is the executive producer. It is remarkable. Mr. McMahon is a singular figure in the world. Certainly not to be glorified, but one to try and understand.

David: Oh man, it’d be fun to do a WWE episode someday.

Ben: Yes. My second quick one is the Dwarkesh podcast. I love the Dwarkesh podcast. I also love Dwarkesh, and I think that if you like this show, you’ll love listening to the interviews he does. Most recent one, or maybe it was a couple ago, is with Daniel Yergin who is the author of The Prize, which is a book, David.

You and I almost read a whole bunch of it for Standard Oil. Then we realized the Standard Oil is over at least the chapter of Standard Oil that we were covering within the first two chapters of his book. So it’s basically everything from the end of our Standard Oil episode forward on the geopolitics of oil that ended up shaping and forming our world today. Dwarkesh is just an amazing interviewer and conversationalist.

All right, well with that listeners, a huge thank you to our partners, JP Morgan Payments, Crusoe, Statsig, and Huntress. You can click the link in the show notes to learn more. We talked to a ton of people for research on this one and while we can’t mention everyone, in part because the list would just be too long, in other part, some folks asked not to be thanked.

We do have some specific ones that we want to give a shout out to. Alex Schultz, the CMO head of Growth, Analytics, Internationalization, great to talk with; Boz, Andrew Bosworth, the CTO; Steven Levy who wrote the book Facebook: The Inside Story, was generous with his time; Jim Breyer, who led Accel’s series A investment in Facebook; Yann LeCun, met as Chief AI scientist; Alex Heath at The Verge for spending his time with me.

To a friend of the show, Arvind Navaratnam from Worldly Partners, who wrote an excellent research report, chronicling everything, well it’s almost like a written version of this podcast. It’s a hundred page PDF that was awesome to consume to help me remember all the big beats of the story, and that his research report is linked in the show notes.

To Arielle Zuckerberg, Mark’s sister; Sheryl Sandberg, obviously longtime COO; Mike Schroepfer, the former CTO and now senior fellow; to Pete Hunt, early engineer who transferred from Facebook to Instagram post acquisition; Naomi Gleit who we talked about founding member of the growth team and the longest employee.

David: Longest-tenured Meta employee at this point besides Mark.

Ben: Yup. To Mike Vernal, former Meta VP of Product and Engineering and former Sequoia partner; Vijaye Raji, former engineer and VP, now of course CEO of Statsig; Aparna Ramani, a VP and AI data and developer infrastructure at Meta, to Owen Van Natta, Facebook’s early COO. David, I know you have a few as well.

David: Yeah, a few for me. To Adam D’Angelo, obviously Facebook’s first CTO; Dan Rose, our early partnerships at Facebook. A big, big final thank you that we owe to Chris Cox, Meta’s Chief Product Officer and leader of the entire family of apps over there.

Really this whole past couple of months at Acquired would not have happened without Chris cold-emailing us, what about a year ago maybe? And saying, hey guys, I love the Nintendo episodes. Because of that, we met Chris, and because of Chris, Mark joined us on stage at Chase Center, and now here we are doing this episode. Thank you Chris for making it all happen.

Ben: Essentially, this episode came about because we had done too much research for the Mark interview and we were like, we probably should do the actual Meta episode too.

Listeners, it is time for our Acquired annual survey. If you have 3–5 minutes, please click the link in the show notes or go to acquired.fm/survey. You might win Meta Ray-Bans, you might win some ACQ dad hats. This is our one big ask of the year and it really, really helps make the show better to hear your suggestions, feedback, and to help show sponsors just how impactful the Acquired audience is. That is acquired.fm/survey.

Check out ACQ2 in any podcast player. If you like this episode, listen to our Nvidia series. Listen to our Microsoft series. I don’t know, maybe go listen to our Standard Oil series. A lot of great Acquired in the back catalog, and discuss it with us in the Slack, acquired.fm/slack.

With that listeners, we’ll see you next time.

David: We’ll see you next time.