Meta's Adam Mosseri explains how Instagram really works - and how he wants to build Threads
01:02:46 Share

This message comes from Rinse. These days, you can do a lot from your phone. Book a vacation, trade stocks, but you can also make your dirty laundry disappear and then reappear washed and folded with Rinse.

Schedule a pickup with the Rinse app and before you know it, your clothes are back, folded, and ready to wear. They even do dry cleaning. Sign up now and get $20 off your first order at Rinse.com. That's R-I-N-S-E dot com.

To remind you that 60% of sales on Amazon come from independent sellers, here's Tracy from Lilies of Charleston. Hi, y'all. We make barbecue sauce, hot sauce, and specialty popcorn. They get help from Amazon to grow their small business faster. They handle all our shipping and logistics, which is a big help. All on it up. Have a great day, Tracy. Hot stuff, Tracy. Ooh, honey. Shop small business on Amazon. ♪♪

From the Vox Media Podcast Network, this is Channels. Peter Kafka, that is me. I'm also the Chief Correspondent of Business Insider. And today we've got a chat with one of the most important people in social media. Actually, he's one of the most important people in tech.

That's Adam Mosseri. He runs Instagram, which is increasingly what makes Meta run, so it's a big job. Mosseri also runs Threads, that is Meta's Twitter ex-competitor slash clone. And when Threads launched in 2023, I was working on a podcast series about the history of Twitter.

And I was really fascinated about how many of Twitter's features and ideas and directions were either something that happened by accident or because users wanted them, or both. And I was really curious how Mosseri, who was able to watch Twitter and every other social platform before building his new one, thought about how threads would evolve. So we talked about that.

And we also talked about more stuff because Adam Mosseri has a big job. We talked about the big pro-Trump pivot that Mark Zuckerberg made at the beginning of the year and how that's playing out. And we got deep on a bunch of questions about how Instagram actually works, like its recommendation algorithms and why it's become so interested in messaging. Okay, there is a lot here. Let's get into it. Here's me talking to Adam Mosseri.

I'm here with Adam Osseri. He's the CEO of Instagram. Are you also the CEO of Threads? Is that a separate title? I don't know if I have an official title, but I like to say I support Threads. You're the boss of lots of big, important social—one really big, important social network and one that's growing. And the backstory here is that in 2023, I pitched you an interview idea where I would talk to you about how to build Threads, how you build a brand-new social network from the ground up. I still want to have that conversation.

Things have happened since. A few things. I want to talk to you about some of those. One of them is that you spent some time testifying in a federal courthouse. I think it was last month. This is the antitrust trial. I'm not going to ask you if you think Instagram and WhatsApp should be split off from Facebook because what are you going to tell me? But you did have – there was some interesting testimony there. You talked about – there were some interesting emails that got surfaced. One was about you talking about the founders of Instagram. Yes.

They were complaining about not being able to accomplish what they wanted to do within Facebook, and then they subsequently left. Given that you were able to sort of watch their discontent running Instagram in a much bigger organization, how has that influenced the way you run the company?

So Mike and Kevin are both really amazing. One of the reasons why I joined Instagram to run product management there was I wanted to work with those two guys. They're brilliant. They built something amazing. They are some of the most insightful product thinkers that I've had the privilege of working with, even though I only got to work with them directly for about half a year.

For me, it was an interesting experience. And if you read the deposition, you'll see, and some of the materials that were at least opened as part of the case, you'll see I'm kind of caught in the middle and trying to mediate as best I can. But I came from... The middle being between them and Mark Zuckerberg. Yeah. And between Instagram and the broader company as well. But I came from the Facebook app, which was...

the only app to start at Meta. And I was running News Feed there, but I was used to having very, very little autonomy. There were so many cooks in the kitchen because there was so much scrutiny internally on what we're doing, not just externally. And for Mike and Kevin, I think one of the biggest challenges is they were going through this transition from being more or less

getting a lot of support from the company, but with the exception of maybe one decision a year, really running Instagram the way they wanted to. Pretty extraordinary for a company to spend dollars and say, go run it the way you want to run it. Six and a half years. Six and a half years they did it. But they basically went from more or less full autonomy to not. And I went from what I thought was basically no autonomy to plenty of it. So for me, I was so excited to have as much space as I could

thought we had. And for them, it was binary in the other direction. And I understand that and I can be sympathetic to that. So I think that you asked what are lessons that I take from it. For me, it wasn't as difficult because I had come from so much less space to make my own decisions to so much more space that I just wanted to appreciate it as best I could. As Instagram becomes more and more important to the parent company,

Do you lose more and more autonomy? It would seem pretty straightforward if that was the case. It's not a linear relationship. So around 2018 is when we really went from operating with a lot of support, but mostly autonomously, to coordinating more across the company. But it hasn't been, it hasn't gotten better.

How do I explain this? There is not more and more oversight every year. There are years where we work really closely together. 2020 was one of those years because there was a pandemic and we were trying to react to the world as it was changing around us. But generally speaking, I still feel like we enjoy a fair amount of space because there's just a lot of things that better does. And Mark and the central teams can't focus on all of those things just because there's only so many hours in the day.

I'm assuming the, I'm going to call it a pivot. You can call it what you want. The thing that Mark announced in January after the election saying, making really big changes in the way we moderate content and lots of other approaches, even down to like how the bathrooms work at Meta. When were you read into that? When did he come to you and say, this is what I'm thinking about doing?

So for central policy issues like those, they are run from, you know, the head of policy and the central integrity teams, which support not only the Instagram and Facebook teams, but all the app teams. So I was read in pretty early on that, you know, a few months before we rolled those out in January. Is there Mark coming to you and saying, this is what I'm going to do? It was a group of us. Mark was pretty clear about him feeling like we were making too many mistakes, him thinking that there was an opportunity to

shift and focus more on voice than we had recently in recent years and that we were going to do that at the company level but we have a small group of executives at meta who meet at least once or twice a week that are the senior most group and so we talked about it in forums like that and and i'm assuming that was a not what do you guys think it's i'm doing this you're going to get on board

The direction was clear. The specifics were not. And I think that's what we worked out to a certain degree as a group, but a smaller group of the people whose primary job it is to define and enforce our policies. We're obviously much more involved than the rest of us. One of the interesting tensions being an app lead at Meta is you...

often are the face of some of these changes, even if you are only playing a supporting role. Sometimes you're playing a leading role, sometimes you're playing a supporting role, but at the end of the day, Instagram is not a company. It's part of Meta, and Meta is the only company that Instagram is a part of. Yeah.

We try to balance those equities. Was there any discussion that, you know, some of these policies might apply to Facebook, but not WhatsApp, or WhatsApp, but not Instagram, and that there could be some sort of like, we're going to handle speech one way on Instagram, but differently on Facebook, or did it have to be sort of universal? We have a strong bias towards being consistent across apps because it's much more difficult to enforce consistently when you have more complicated rules that

change across surfaces. Now that said, there are some pretty big differences between how we enforce in public contexts like Facebook feed and Instagram feed and private contexts like messages, whether it's on Messenger or Instagram or WhatsApp. But that's less to do with the app and more to do with I think people expect a fair amount more privacy in a context like messaging.

And one of the recurrent – it's almost kind of amusing to me, probably less so for you, is watching you handle usually pretty angry questions about political content on the site. Never. It never happens. Obviously, there was a lot of – there used to be a lot of politics on Facebook and I guess Instagram to a lesser degree. Then you guys pivoted away from that and now you're saying it's back again.

You were asked about it in court and you've talked about it publicly, but I'm still a little confused about where we're at today. I can more or less type whatever I want about anything political and that sort of extends culturally, right? Correct. And then does my content get treated differently based on – will my reach change or is it fully neutral? So –

At this point, your reach won't change. The unfortunate complexity is the way we tried to not...

amplify politics was through not all of the experiences on Instagram and Facebook, but specifically recommendations. So content you were seeing from accounts you have not chosen to follow. So on Instagram, always, if you posted political content and people decided to follow you, we ranked that content the same as non-political content to the people who have decided to follow you.

But we did not or we tried to avoid recommending proactively content to people from accounts that they did not follow about politics. That's the big change. I was pretty public about this. It's not that politics can't drive some engagement. It definitely does.

It just drives an immense amount of scrutiny and criticism and often a reputational and a brand hit that comes along with it that I don't personally always find worth that trade-off. So given your apprehension about that, you guys rolled this out in January. How's it worked in the real world? Are the things that you were concerned about, are those playing out?

So the amount of politics on all of the apps has grown, but for the apps that I focus on, Instagram and Threads, it is still a relatively small percentage of the content as best we can tell on Instagram. It is a large percentage of the content on Threads. Threads is about sharing perspectives and politics are...

top of mind for a lot of people right now. Whereas Instagram is more about sharing creativity, photos and videos. And yes, some of those can be about politics, but most aren't. So it's affected threads a lot more than it's affected Instagram. Do you draw distinctions between someone saying, I like Democrats or I like Republicans versus I'm interested in wellness and I think that RFK Jr.'s plan to remove fluoride is a good plan. Is that a politics post or is it a politics post if you say...

You can see where I'm going with this. Especially in wellness, there's just a lot of stuff that... But this could be, I think, a million different topics where it can be read as a political statement or it is a political statement, even though it might not be traditionally considered one. So this was the challenge with trying to do anything different for politics historically, was that drawing a bright red line around what is and what is not politics is really difficult and a borderline fool's errand. So one of the reasons why...

I didn't put up as much resistance as I might have a year or two before is our ability to actually consistently and accurately identify politics. You were getting tripped up on that stuff. We were just, we were missing a lot of it and we're getting it wrong a lot of the time. And so...

A policy is only as valuable as your ability to enforce it. And we were making tons of mistakes. There was that anecdote about Mark Zuckerberg being frustrated because he'd posted a photo about recovering from a surgery. Yeah. And that was interpreted as a health-related thing. That got caught up in some health safety measure on the Facebook app, not on the Instagram app. And, you know, things like that, you know, mistakes happen. And we track how often they happen. But if they're happening tens of percents of the time, then that's far too much.

So in your specific example, based on where the policies were, yes, those both would be politics because you were talking about a specific opinion about a political figure today. But... What if I just took out RFK Jr. and said, get rid of fluoride or fluoride's bad for you? That technically wouldn't be, but we could have very easily made a mistake there. And so now that politics are back and it's more at threads than other places, but...

What's the reaction? How are you gauging reaction from users? People are posting it more. Are there people who enjoy that? Are there people who say, I really didn't want this. I want to turn away from it. I mean, I think that was initially one of the reasons you guys were pushing it away, right? You thought that it was turning users off in addition to inviting scrutiny. Yes. It's polarizing. So certainly some people are enjoying it and others are not. The thing that is difficult is

With any ranking system, like we have on Threads, you rank for what you can measure, and it's easier to measure people's first-order preferences than their second-order preferences. So, you know, do I immediately have a reaction to this post? Is it easier to measure than would I say this was worth my time if you asked me tomorrow?

And we bias towards those first order preferences because they're easier to measure. And so we rank for them more directly, not because second order preferences matter to us any less, but because they're more difficult for us to understand. And because of that bias, my take is that politics often are overvalued by ranking systems because we think people value them more than they do. Because the way we measure if they value them is,

Did they like it? Did they share it? Did they click on it? Did they reply to it? Did they reply to it? And those do correlate with thinking something was worth your time, but they over-correlate, let's say, in the context of politics. Because I might be sharing it with someone because I think it's a great idea, or I'd be saying, look at this idiot.

You think you should get rid of Florida. Yeah, or even if you don't even value replies or shares, you just have a strong reaction. You're more likely to like a political opinion you agree with when it comes by in your feed than...

piece of news about what's happening in the entertainment world that might matter more to you as a person, but doesn't invoke the same emotional immediate reaction. Have you guys discussed it saying, look, six months into this, a year into this, let's have a big sit down and talk about what's really happening on all the platforms with these new policies and maybe we can amend them, roll some of them back, or is this sort of we're going down this road and we're not really going to deviate?

I think how we go about it, we will iterate on. So if you wanted to rank political content differently, not to suppress it, but to just get better at connecting people with the content who are interested in that content, you would still need to be able to classify political content with high precision and recall, which are technical terms we use for the ability to find it all and to accurately find it. And so we're still working on that because it might turn out that...

We no longer try and suppress politics in recommendations, but how we rank politics might be different because sometimes how we rank different topics is different because different topics sometimes merit different approaches. It seems like a distinction without a difference. If it's

If you're talking about the ranking versus how, whether it's suppressed, I mean, it seems like you're using different words to describe how often I'm going to see something one way or the other. No, I think it would be a difference. So for instance, let's say it turns out that, so we can do things like we can run surveys. We do this. We ask for certain posts, how worth your time was this post? And then let's say it turns out that we do that, now I'm being hypothetical, and there is a large correlation between whether or not you liked something and whether or not you said it was worth your time.

But for political content, there's no correlation. Or more likely clicks. You know, clicks often have very little correlation with whether or not you said something was worth your time. You could just value clicks less for political content and value other things more. The amount of political content in the system overall would be the same, but you would be ranking more effectively.

So in terms of what you show me personally based on what I'm telling you. Yeah. Well, you, but also in aggregate, maybe some people would see more and other people would see less. For people who are clicking on political content in this hypothetical example and don't like it, they would then see less. And for people who are liking political content in this hypothetical example and that got more value in the value model, they would see more. And the overall amount of political content would be about the same.

Someone smart told me once that a major difference between the way TikTok works and the way Reels works is there's just so much more data collected by TikTok. They've got many, many more signals that help them figure out what's going on. And two, they are very willing to show you something with basically zero likes that no one's seen just as a sort of a baseline to sort of round things out. And you guys are more reluctant to do that. Is that accurate?

Partially. I do think they are more aggressive on data in general. But the latter part is, I think, where there are more differences and those differences are shrinking over time. But when you are trying to show people things that they are interested in, an easier way to do that is to show them content that you have lots of data on because lots of people have already seen it. Yep.

That in the world of ranking sounds negative. In the world of ranking, it's actually more of an objective term. But that's called exploitation-based ranking. So you're exploiting the data that you do have. Not people. You know something about this post. A lot of people like The Rock. You can show someone else The Rock with high confidence that they might like The Rock. A lot of people like looking at pictures of women in bikinis. Here's more of that. Sure.

With some personalization, not just like blanket. Now, what TikTok does really well and we're getting better at and YouTube is getting better at and others are, is called more exploration-based ranking. So looking for ways to make sure smaller creators and content with very little, very few views gets a chance. And they do that in a variety of ways. And we've been...

catching up, but they're still better than us. Is the impulse there, I want to help this creator who needs to break through, otherwise it's just the same X number of people with an enormous distribution? Or is it, I want the data from this random person who's giving me something we haven't seen before? I guess both things can exist at the same time, but where do you think the ultimate impulse is? I think it's more about being a place where there can be upward mobility.

For a bunch of reasons. One, I think you'll end up being more culturally relevant because some of the most interesting things that happen in culture happen with smaller names. Two, because it's harder to pull big creators from other platforms than it is to grow a new generation of culturally relevant creators in your platform. And they'll be more loyal to you because they grew up on your platform.

So, and it's also just the promise of the internet, right? Is that we don't all have to listen to the same 40 radio stations. You might be into Afrobeats and I might be into Italian jazz. And we wouldn't have known that 50 years ago, but we can figure that out now and then get connected to that kind of content. So the way it works is essentially almost an audition or a cascading audition. So you try to get every post that gets posted by a creator online.

call it a thousand views. I mean, I'm making numbers up now. And you do that by setting aside some percentage of impressions and showing this content even though it's not the most interesting content for that individual person. Then you take those, call it in this hypothetical example, a hundred reels that you got a thousand views each. Then you take the top 20% of those that did the best and then you do that again in another pool with another thousand, with 10,000 views each.

And then you take the top 20% of that and you do another pool. So you basically can win this informal, invisible competition and get more and more views and engagement. And that's how you see things go viral from accounts that are incredibly small on all the platforms now. Now, at this point, we do this type of ranking, exploration-based ranking. TikTok was, I think, the first to really do that at scale. YouTube does the same.

But that is about trying to find gems that people might not know about yet. Be right back with Adam Mosseri. The first word from a sponsor. In business, they say you can have better, cheaper, or faster, but you only get to pick two. What if you could have all three at the same time? That's exactly what some of the world's most innovative brands and AI tech companies have, since they upgraded to the next generation of the cloud, Oracle Cloud Infrastructure.

OCI is the blazing fast platform for your infrastructure, database, application development, and AI needs, where you can run any workload in a high availability, consistently high performance environment, and spend less than you would with other clouds. How is it faster? OCI's block storage gives you more operations per second.

Cheaper? OCI costs up to 50% less for compute, 70% less for storage, and 80% less for networking. Better? In test after test, OCI customers report lower latency and higher bandwidth versus other clouds. This is the cloud built for AI and all your biggest workloads. Right now with zero commitment, try OCI for free. Head to oracle.com/vox. That's oracle.com/vox.

Avoiding your unfinished home projects because you're not sure where to start? Thumbtack knows homes, so you don't have to. Don't know the difference between matte paint finish and satin? Or what that clunking sound from your dryer is? With Thumbtack, you don't have to be a home pro. You just have to hire one. You can hire top-rated pros, see price estimates, and read reviews all on the app. Download today. Support for this show comes from Pure Leaf Iced Tea.

You know that point in the afternoon when you just hit a wall? You don't have time for self-care rituals or getting some fresh air, so maybe you grab a beverage to bring you back. But somehow it doesn't do the trick, or it leaves you feeling even worse. What you need is a quality break, a tea break. And you can do that with pure leaf iced tea, real brewed tea made in a variety of bold and refreshing flavors with just the right amount of naturally occurring caffeine.

And we're back.

While we're on algorithms, I'm old enough to remember getting Facebook news feed and it was a chronological list of things people I knew posted. First launch was ranked. We changed it to chronological in March of 2009. I was the designer on that. All right. Then I give you credit for that. Twitter certainly was chronological. And now everyone has shifted to some kind of algorithm.

give you some kind of choice to go back to chronological if you really want to. If you're a nerd like me, you can do it. And there's always a list of complaints from people like me saying, why am I getting algorithmic stuff? I just want to follow the people I want. I've heard different answers, but I want to hear your answer. Because it's not just you, it's literally everyone who runs a platform. Why does everyone who makes a social product think that I'm better off seeing an algorithmically curated feed versus one that I've put together for myself that's chronological? And it's more than just...

It's also now more and more recommendations, content from accounts you haven't even decided to follow yet. It's because it's the only way to grow these experiences. So the amount of content people post publicly in feeds is, generally speaking, going down across the entire industry.

Because people are moving more and more sharing to stories, which you could argue is a different kind of feed, but even more into messaging, group chats, one-on-one chats. On Instagram, there are way more photos and videos shared into DMs, no texts, put texts aside. More photos and videos shared into DMs than into stories, and way more photos and videos shared into stories than into feed. So if the amount of content you have to rank is decreasing, the overall how engaging feed is, is also just decreasing.

It's just getting worse. And so we rank to try to make the most of your time, show you the stuff that you're the most, we think you're the most interested in. Again, I educated a guess. But we show recommendations because you might follow 200 accounts and one in 10 of them posted. So we've got 20 things and we can reorder those 20 things, you know, 20 factorial ways. That's only so much upside. Whereas if we look at the billion things posted in a given day and we find something you're interested in, there's more upside. Yeah.

Is there a point where someone has to go, well, wait, if people are going to these apps, to these social networks, and there's less stuff being posted there, shouldn't that be a signal to us that we're doing something wrong? Like that people don't really want the experience that we've been serving up to them? I don't think so, because I think there's a difference between the consumer and the producer's incentives. So...

There are more and more creators out there producing more and more content, but those who are sharing to feed are more and more creators and less and less average folk like you and me, or maybe you and me 10 or 15 years ago. But that doesn't mean that the average consumer doesn't want to spend some time learning about the world or being entertained or connecting with somebody that they think is interesting. You can think of it as a two-sided marketplace. There's producers and consumers, and you're trying to facilitate a valuable exchange.

This push towards messaging, you've been talking about it for a while. I'm assuming that you are arguing that this is something consumers, users were doing sort of on their own and you guys are responding to it.

Oh, yeah. It's a paradigm shift. It's something that's happening to us. And I've heard that, you know, people are going to group chats because they... It's usually brought up in the context of they feel like it's safer, they have more candor, whatever it is. But it's also hard for me to imagine lots and lots of people who aren't people like you and me thinking about this stuff all the time are literally thinking about how their posts are going to be received. Is there some other reason people are sharing more privately versus publicly? I think...

I mean, so it's not most, most of it's not group chats. Most of it's one-on-one chats. Right, right. I think the foundational reason is that there are more things that you would feel comfortable saying to somebody

one-on-one than things you would feel comfortable. So it's a little bit that simple versus I'm going to tell you something versus I'm going to shout on the street. Basically, yeah. I mean, this is a weirdly sad example, but you could think of sharing in feet as standing on top of your roof and yelling something at a hundred people and hoping that 20 people hear it. Yep. There's some things I would do that for. I'll do that. You know, my kid hits his first, you know, baseball or, you know, I get a job. Your kid's very cute. Here's a picture of him.

But the average thing, the amount of things I would say to you on a phone call, my wife on a phone call, my best friend on a phone call, there's a lot more of those things. I think that's the most important reason. How does that shift affect the business of meta if people are sharing privately? It moves more and more of that friend content into private experiences. And then the question is,

Can you either make those private sharing experiences symbiotic with the ones that we monetize like feed and stories? Or can you monetize those experiences directly? For Instagram, the thing that has been amazing is that we have leaned into video and reels in a way that actually grows messaging. So historically, when I've worked on the Facebook app,

We leaned a lot into video in 2014, 15, 16. We were very focused on trying to catch up with YouTube. And growing video grew the amount of time spent in the Facebook app, but it decreased everything else. It decreased messages, comments, likes, revenue, because there's less ads per minute or hour of time. And with reels on Instagram, because they're short and because they're entertaining and because we rank them often for what they're going to, who's going to send it to a friend,

People see these, you know, I'll see a stand-up comic doing a bit that I love and I'll send it to my brother because I know he's going to enjoy it. Or I'll see a piece of politics and I'll send it to you because I think you might be interested in it. And then you and I talk, maybe you look at your feed, maybe you engage with something else, maybe you send that to somebody else.

So this private messaging part of the experience, we've managed to build it in a way that's very symbiotic with the public context like feed and stories and reels, which we monetize directly with ads. So we're going to show you engaging stuff, this algorithm. You're going to engage in it, and we'll be able to monetize your eyeballs like we always have. And then you'll share it with other people, and that will then...

Create a virtuous cycle, basically. They'll stay more engaged. It's a positive feedback loop. And it's, I think, important, particularly for Instagram, because...

We are about connecting with your friends over creative things. I mean, for some people, we might be a pure entertainment-based or public content-based app. But we want friend content to continue to be a core part of the experience for most users. And this allows, in some ways, Instagram to stay social but still grow as a business.

I could spend hours asking you more questions. Some of them I've asked you before, so I'm not going to keep asking you about monetization. Although I still want to know when you guys are going to share revenue, but you won't answer that, so that's fine. I'll answer. I'll explain that. But we should go and we'll come back. Let's talk about threads. So I didn't realize until I guess this came out of the trial that originally threads was supposed to be an Instagram feature. So we were talking about different ways to compete more directly with Twitter and

Can I stop you there? Yes. Because why? Like, I know that Twitter and Facebook in 2010, right, fiercely competitive, big deal. Twitter was a place for real time. And then basically that competition stopped because you guys just lapped Twitter like over and over and you won, right? And there were many more people who wanted to engage in a Facebook and Instagram-like experience than they did on Twitter. So why bother going back to Twitter?

When I joined in 2008, it was one of our biggest fears. That's why we did the chronological feed in '09.

For a number of reasons, but I think Twitter is a great app in a lot of ways. I use Twitter a lot still. X, we'll call it. I think it's better for public conversations. I think there's a lot of, though it's not the biggest social app, there's a lot of cultural relevance. There's a lot of really vibrant, amazing communities on X. There's, you know, the NBA Twitter. There's Black Twitter. There, you know, there's these insular networks kind of like VC Twitter and crypto Twitter. And, yeah.

Part of what we care about at Instagram is being a place where creatives do their thing. And I think X does a good job there.

So your thought was we would maybe like to have some kind of Twitter-like experience. And the initial thought was let's bolt it on to integrate it into Instagram. There were a number of different things there. So we were, that was around the time we really accelerated our work on broadcast channels on WhatsApp and on Instagram and on Messenger, which by the way are a big deal in a lot of the rest of the world, particularly popularized by Telegram.

We looked at and had a bunch of designs for building something like Threads as a tab into Instagram. And we did consider and ended up building a separate app. And there were a lot of contentious debates. I was living in Europe at the time. So for me, very late night contentious debates about what made the most sense. Where did you want what's now called Threads to live? I was excited about channels, but Mark made the point and I agreed with him that channels are going to be

They're not going to be a place where you keep up with tons and tons of culturally relevant people. They're going to be a place where you subscribe to like the five or 10 you care about most. I was more bullish on something within Instagram. The two issues were one, my concern was trying to build a reply model, which I think Twitter gets a lot of credit for popularizing into an app with comments. It was just going to be confusing. So I never really had a good answer to that.

And then Mark's point was an app, a separate app will be harder, but if it was successful, it would be a more valuable thing to create in the world. And he, a lot of what Mark does is anchor us really high and said, no matter how strong a year we have, the question is always how we can do better. And he, what was going on? It was late. I was in Italy for my anniversary with my wife and,

And he's like, well, if you were going to do something bigger, what would you do? So I was riffing and I kind of pitched a version of threads. Said we'll lean on Instagram, strength with creators. We'll use Instagram identity. You can bootstrap it with their graph, but we'll focus on basic replies and threads. And I called it Textagram as a joke, which unfortunately stuck as a name for months before I managed to kill it.

And Mark's like, yeah, that's a good idea. We should do that. And I was like, I don't think we should do that. And he said, well, don't put someone in the boss unless you want to do it. In the classic Mark move, he said, okay, but if you don't do it, I'll have somebody else do it and it'll be built on Instagram. And I was like, okay.

Sounds like I'm signed up. So he gets the credit. So this is where, this is basically sort of what I initially wanted to talk to you. So you say we want to create Textagram. Great name. Don't say it again. And you said we want this to be Twitter, right? We have our version of Twitter. And you've been able to watch Twitter for a decade plus at this point. And not only that, every other social network, the ones you run, everybody else's,

What did you want to replicate? What did you want to go, we definitely don't want this to be replicated? We knew it was going to be a place where a small number of people created most of the content that was consumed. Feed apps are always that way. Text is always that way. This is a text feed app. So that was okay. We wanted to embrace that. We wanted it to be a place where the people who had the most relevance in their worlds and their industries would share their perspectives, their hot takes.

and have ideally conversations about those with other people of influence. And we just wanted it to be less angry. And that was the core. Those were the core ideas. We then added more. I mean, we leaned on Instagram to bootstrap the experience and the app. And we also got excited about being open and integrating with the ActivityPub protocol on the Mastodon servers. But the original kernel of the idea was

a place to support all perspectives. So there's like product stuff I want to ask you about, but that main idea that Twitter is an angry place, right? People who love Twitter would still say, yeah, it's an angry place and I use Twitter still way too much even though I've cut way back and I'll get really engaged and really unhappy using it. I'll be really angry at someone.

But is that a Twitter issue or is that a text issue or is that just how people interact on the internet? If they're going to talk about ideas, they're inevitably going to get angry. It's all of them. I mean, look, the internet, there are billions of people on the internet. There are billions of people on Instagram. Some of them are going to be nasty and angry and they're not going to check that at the door when they open up the app. So there's no version of...

a open social network that is all positive and I don't think that was what the intent was but I do think there are things you can do to change the tone and I think you see that I think Instagram well what you rank for matters a lot so you're ranking for replies or any replies are you trying to identify some form of more constructive conversation which is harder to do

But that matters a ton. I think one of the reasons why Instagram is more positive than Facebook and Twitter is because it leans more into imagery and less on text, which obviously we couldn't do with threads. But I think that also lends itself to a different place. There are tools you can build. And we've done a lot for threads. So things that allow you to, you know, when someone says something really nasty as a reply to one of my posts on threads, I can hide that.

post from everybody, not just from me. You can do the work manual, yeah. Yeah, and people do a lot. Because if it's a small number of creators creating most of the content, if you can build some powerful tools for them to decide if they can be quote posted, to remove quote posts, to decide to hide replies, decide who can reply to them, these things can go a long way. The hidden words is another one. So this is on the third bucket of tools to allow you to... We know the hidden words meaning...

You can block certain words. Individuals, individual creators and or users can say, I don't want to see. Yes. In my feet. Yeah. Or like, you know, I get brigaded once a month because some creator somewhere got too many strikes and got disabled. And then they ask another creator to point everybody at me. And so sometimes I have to just block that name from all of my comments and all my posts.

because otherwise I get thousands of angry people. And I often look into it and it's usually, I can't reverse it, this person posted one too many violent videos or whatever it is. But the fourth, just to finish the four real quick before you jump in, I also think it matters what you say you want the app to be. I think the way we positioned it as trying to be a friendlier place meant there was a selection bias in who signed up and who stuck around. And that is not true.

I'm not even sure it's possible to measure that effect, but I believe that effect is real. From the outside, one of the things I think I encounter a lot, and other people talk about it as well on threads, is a user coming in from Instagram or Facebook.

and not sort of fully understanding what it means when their thing goes to threads and maybe something that would have made sense in an Instagram post or a Facebook post and it shows up in Textagram. You're making me regret telling you this. Every podcast we do, you're going to regret one thing. That's the kind of thing about podcasts. Yeah, and it just doesn't make sense. And often those posts go viral because it's like someone wandering into a bar

wearing, I don't know, a leotard because they were going to go to dance class, but they're at a bar instead. Yeah. Metaphor, but you get it. I get it. First of all, am I right in thinking that a lot of the user base is coming from Instagram and Facebook?

There is a lot, but the way it works is that there's a core group of people who are proper threads users. They open up threads directly or they get notifications from threads and they open the app up from those. And they're sort of our threads-driven users. And then there's a lot of users who come by from Instagram and from Facebook to check it out. And some percentage of those turn into proper threads users over time. And so there are sort of two

one way to think about it is there's two levels of users. They just seem like entirely different populations though. They, well, the former turns in, sorry, the latter turns into the former, which is part of why the app continues to grow, not just in how many people are using it, but in engagement and another core health metric. So you don't think there's like, there's different kinds of humans and some are

thread slash Twitter humans and some are Instagram humans. No, I mean, I'm not sure that's, I don't know that I would disagree with that, but I think that the core threads users will be made up of a combination of people who leave competition like X because they want to get an alternative and people who never used X because

but actually want a place to read about the perspectives around the industries that they care about. But they come from Instagram and Facebook and elsewhere, but for whatever reason, they never really jumped into Twitter. That's the, yeah, they tried it, didn't like it, or heard about it, or tried it once, and you're sort of giving them another bite at the apple. Yep. Along the same lines, another thing I notice a lot is, and well pointed out publicly, is

Because you guys emphasize replies and threads, oftentimes I'll find a post that has a lot of engagement and it's something stupid that someone published either unintentionally or intentionally. Either way, it doesn't matter. And most of the replies are, you idiot, or some version of that. And that's not good for me. I don't enjoy it. What am I missing here? It seems like I could imagine how that was set up.

In 2023, when you guys were first playing with it, it's surprising that I'm still seeing it in 2025. So interestingly, we actually didn't value replies at all in the ranking model in 2023 when we first launched because we wanted to be careful not to incentivize a bunch of angry replies.

We now do value replies, but not that highly in order to avoid these use cases. I could have sworn Mark Zuckerberg said that was like a primary thing they used to break it. It's one of them, but it's not the most important one. Okay. And it's growing, but I think we're trying to grow it slowly in its importance to try to minimize the number of scenarios like the one you described. It may happen, but they don't happen a ton if you look at the overall percentages. We'll be right back. But first, a word from a sponsor.

Support for this show comes from Smartsheet. You know that look the basketball player gets as they step up to the free throw line? Or the rock star right before they belt out the high note? Or maybe even your dog as she gnaws through your favorite sneaker? That all-encompassing, nothing-else-matters, right-here-right-now feeling has a name. It's

It's called Flow. But you don't have to be a professional athlete or musician or happy dog to experience Flow. You can access a focused flow state in everything you do. And Smartsheet is here to help you do just that. Smartsheet is the work management platform that helps remove friction and distractions, allowing your team to seamlessly take the shot, sing the high note, or complete the big project.

Smartsheet's innovative platform transforms the way teams operate, helping them automate workflows and cut through the chaos of whatever they're working on. No matter the task at hand, that feeling of flow is at your fingertips when you use Smartsheet. Smartsheet. Work with flow. Learn more at smartsheet.com. Support for the show comes from Mercury.

What if banking did more? Because to you, it's more than an invoice. It's your hard work becoming revenue. It's more than a wire. It's payroll for your team. It's more than a deposit. It's landing your fundraise. The truth is, banking can do more.

Mercury brings all the ways you use money into a single product that feels extraordinary to use. Visit mercury.com to join over 200,000 entrepreneurs who use Mercury to do more for their business. Mercury, banking that does more. Support for this show comes from Pure Leaf ICT.

You know that point in the afternoon when you just hit a wall? You don't have time for self-care rituals or getting some fresh air, so maybe you grab a beverage to bring you back. But somehow it doesn't do the trick, or it leaves you feeling even worse. What you need is a quality break, a tea break. And you can do that with pure leaf iced tea, real brewed tea made in a variety of bold and refreshing flavors with just the right amount of naturally occurring caffeine.

With a Pure Leaf iced tea in hand, you'll be left feeling refreshed and revitalized with a new motivation to take on what's next. The next time you need to hit the reset button, grab a Pure Leaf iced tea. Time for a tea break. Time for a Pure Leaf. And we're back.

Did you guys think about the fact that when you roll out a new social network in 2023, there's just a large population of people that didn't exist 10, 15 years ago who say, oh, there's a new social media property. I want to take advantage of it either for lulls or because I want to make money or whatever. And they sort of have expectations about how it's going to work. And they immediately sort of import whatever they were doing on Instagram or name your platform. And they try to move it over there.

We thought about that space. We've seen a number of things like that. So one is checks in place or be less sophisticated in its ability to identify things like scammers and spammers. And so we dealt with a lot of that. And some of it was gnarly, but some of it was just clickbaity stuff. And I think we've managed to get that most under control, but that's still an issue.

Another is not a negative one, but not what we were really going after with Threads, which was brands. You have all these social media managers who are like, oh, I know that there's a first mover's advantage. And so you had this proliferation of really quirky, kind of punchy brand accounts on Threads. Why not? If that works on Twitter, why not have it on Threads? I'm not against it, but I don't think it's as valuable as hearing...

If you're into the Premier League, what trades do you think are going to come up between your favorite teams from your favorite commentator? It's less differentiated content. It's less culturally relevant. It's engaging, but it's filler. It's not sort of a reason to join an app. So we had a lot of these different patterns. People would come over to try it because it was new and because they had a...

hypothesis for some value that they could extract either by replicating what they did elsewhere or taking advantage of how this was different and some of it good and some of it bad and we just try to manage it and minimize the bad and maximize the good. How do you think about people who come over and say this feature existed on Twitter? You should have it here, right? Direct message, you hear it all, but whatever it is, at replies, whatever it is, a thing they had over there and they want it to be here.

How do you think about that? How do you weigh, and this is core, I think, right? Like, how much of this stuff do you want to import? How much of this stuff do you want to say, we've got a better idea here? I think that you want as much of the work you do to be meaningful, defined as either growing the app in some measurable way, or just being interesting and novel and creating a delightful experience. But ideally both.

And most of the fast follow features that people have asked for from Twitter, we've built. And almost none of them have been meaningful by that definition, right? Because they didn't meaningfully change engagement as much as people keep assuming that they will. Some of them even regress engagement. And they're by definition not novel because you're just fast. Well, it's an example of one that people wanted and have regressed engagement.

Oh, a following feed. You can have a feed that is chronological. You can even make it your default and that just progresses engagement. But other things, you know, people are like, oh, add hashtags, doesn't move engagement. Add the edit button, doesn't move engagement. Extend the edit button window, doesn't move engagement. Add feeds of lists, doesn't move engagement. But we do them because we want to make sure that we're listening to people and that we are earning some brand credibility and...

and hoping to, you know, have people love the experience. But it doesn't, those features are all good to build. I don't regret any of them. None of them moved engagement and none of them were interesting because they've all been done before.

A lot of the features that started at Twitter were ad hoc, made by users. The company didn't intend for them to be used that way, or people just built stuff and the company responded, right? The app reply is one of them. Do you want that sort of experience to happen on threads where the users say, you didn't expect us to use the platform this way, but we are, and now you're going to have to adapt to it? Those are great. Those are sort of happy accidents or abuses that become uses.

Like when Twitter was kind of built into the company, for better or for worse, the users were in a lot of ways running that platform. Hashtag was another one, right? Yeah. Where they would, that was an organic thing. So I think that's great. We've got some ideas in the pipe that are responding to patterns that we're seeing in usage, but also some ideas that we have that we're excited about that are newer and more different that hopefully will come out over the next couple months. But we just had a lot of basics. I mean, you also have to remember, when we launched Threads,

There might have been 50 people on it. There were four ranking engineers. No one... In our wildest dreams, we didn't think 100 million people were going to install the app in the first week. Well, that's why I was also confused because when you launched, it seemed pretty clear that this was not a... You know, it was a small project for you guys and there was a lot of like, oh, is he going to take on Twitter and is he going to take on Elon? I'm like...

It's not staffed that way. How big is the staff now? Still small. It's a few hundred people, low hundreds. But we get a lot of leverage. So that tells me that it's not a very important app within the... No, I don't think so. For me, I'm trying to treat it like it's its own app every way I can.

The thing that I think we can do is I actually think we can run an app much more leanly than we have in the past. And that's easier to do if that's in your DNA from the onset than it is to do retroactively.

So, I mean, I spent a lot of my time on it. I put a lot of my best people on it. We've grown it a lot since then. But I do think my hope would be we can run this app with hundreds of people instead of thousands of people, in part because we get an immense amount of leverage out of the rest of the company. If we have this conversation a year from now, what is going to be meaningfully different about Threats?

What success looks like is that it's much more interesting than it is. I think Threads is engaging and it's growing, but I don't think it's as interesting as it could be. A lot of the content is pretty fluffy or commoditized. You can find it elsewhere. So I think what success looks like in a year is you feel, even if you don't use it any more than you use it today, that it's more interesting. That for whatever you're passionate about, you can find some perspectives from people you respect about that passion.

How to get there? We've got some ideas I'm excited about around building communities. Not a giant group that everyone can subscribe to because I think that's always too literal and the most interesting communities aren't usually that structured. But whether it is helping people, us understanding different communities that exist, understanding your affinity for some of those, understanding who are the most influential voices within that community, not because they're the biggest, but because they are the most trusted communities

what's maybe trending in those communities. These are some of the ideas I'm most excited about. Is that taking the users that are there and sort of reshuffling them and saying, oh, you're doing interesting stuff here. Let's elevate you or figure out ways to...

distribute you more widely versus like, I know that for a while you guys were interested in sort of, it seemed like you were interested in, in replicating or competing with sports Twitter and NBA Twitter specifically. And Adrian Wojnowski, when he was still the king of NBA Twitter moved over to threads and he was posting there, but you could tell like he was still doing his most important stuff. The deepest stuff was happening on Twitter. Um,

NBA Twitter, NBA Threads has still not taken off. I keep going to check because I'm a Wolves fan. You don't follow Shams on Threads? I don't, but I get enough of this stuff. But the point is, I know that if he's publishing his most important stuff, it's going to go to Twitter first and it's going to go deeper there. Which Roger didn't necessarily, though. I talked to him about it before. I don't know. I did a lot of compare and contrast. And he would have maybe at the same time, because I'm assuming you were incenting him to do that, but it was still a deeper experience on Twitter. You could tell that he was still a Twitter creature.

So do you have to find, do you have to convince those guys to become threads creatures or do you have to find organic threads users who love the NBA and promote them? I think you have to do both. I think that growing a new group of influential creators around a sport like basketball is going to be necessary. Otherwise, you're not going to pull everybody over, but you also have to pull some people over as well.

So, I mean, we'll take the NBA threads as an example. It's growing. It is not as big and as vibrant as NBA Twitter, but it is definitely growing and it has its own set of characters like Mio Rush and others. And Shams does post immediately on threads as well as X. All these folk tend to post everywhere at the same time. Sometimes they use an app that is a meta, not a meta, but meta with a lowercase m to post everywhere.

But I think if we wanted to do that, we would have to do, but we'd have to both attract more talent from Twitter and build up more homegrown talent, which we are also doing. And I think that an interesting way to do that would be, because we don't do this well now, is if you, we should know that you're into basketball and we should help you find a bunch of really interesting basketball accounts that are on threads that you don't even know are on threads or know about, and I'm sure they exist.

And then we should help you and see what those people are saying and what's trending within that world. And those are things that we don't do very well right now. And do you feel like you've got to lean more into the real-time nature if you're going to make...

or others. It looks like you're ramping that up around specific events where you sort of know something's big. You guys are pushing more to get that in my feed more often. The thing I hate the most about threads is finding a post from three days ago. Yeah, no, search is not very good. I'm sorry. No, no, no. It's just showing up in my feed. Oh, you're getting the bumps. That's not good. I'm just getting the, you know, this was, it's a standard joke we all make, you know, in two days, this will be a funny post to read. Yeah, yeah. No, that is a, that is a thing.

So threads is ranked much more for real time than Instagram or Facebook. So the average age of a post you see is way shorter. There is a perception that there's a lot of old content. It's actually a very small percentage of content, but it's very noticeable when you see it. And that's part of where the meme comes from. So we have to get a handle on that.

That, by the way, can be okay. There can be content from two days ago that is great. But if it's what happened in the third quarter and in the fourth quarter everything changed, it's completely useless. So we have to get better at differentiating between those two. But it is much more focused on real time than any of the other apps we work on. Okay. I'm going to let you go. But before I go, speaking of saying things you regret, the last time we talked was 2021. Yeah.

It's right after the journal had written a bunch of really scathing pieces about safety on Instagram and Facebook in general. You had a quote. You said, we know that more people die than would otherwise because of car accidents, but by and large, cars create way more value in the world than they destroy. I think social media is similar. I 100% understood what you meant. And more than that, I know that lots of people in tech, at least

used to, and I think still feel that way, you've regretted that comment. Did you regret saying it out loud or did you regret the actual thought behind it? I think a better analogy would have been something like magazines. That didn't involve people dying. Yeah, I think I was tasteless in retrospect. I think that those of us who are lucky enough to be running some of these large platforms have an immense amount of responsibility. I think that

We all need to, and I hope I do, take that responsibility very seriously. But I also feel like it's important that we're honest about the fact that given there are billions of people on these platforms, there are going to be good things and bad things that happen on the platforms. And we need to maximize the good and we need to minimize the bad and we need to be transparent about that.

But I always try and be forthcoming about the fact that, yeah, with billions of people on a platform, there are going to be some negative things that happen and that number will never be zero. I will always work to push it as close to zero as I can, but it'll never be zero. But I also feel like

I decided 10 years ago to start engaging on Twitter and engaging with some of our biggest critics. And I know then, as I know now, that I will say things and get slapped around for them. But the conversation is going to happen with or without us. And I'd rather participate in that and try to do so with integrity and honesty, even if I get slapped around for saying something not as thoughtfully or carefully as I might have.

All right. I think you'll be fine on this podcast. I don't think anything is going viral from this one. Something is going to happen every year. It's okay. I mean, it's usually now it's usually more of a meme. It's usually like me getting made fun of by The Daily Show or somebody like that. I'm going to try to work on a meme for you. Adam Mosseri, thanks for your time. Thank you for your time, Peter. Appreciate you.

Thanks again to Adam Mosseri for coming on the show, saying things he probably won't regret later on. Thanks to Travis, who produced and edited this show. Thanks to our advertisers who bring it to you for free. And thanks to you guys for listening and writing. And you should share this. Share this on Instagram if you want. I don't care however you share it. Just share it. We'll see you next week.

McCrispy strips are now at McDonald's. Tender, juicy, and its own sauce. Would you look at that? Well, you can't see it, but trust me, it looks delicious. New McCrispy strips, now at McDonald's. With HubSpot's suite of AI-powered tools, you can get more done way faster. Speed up your lead generation and create attention-grabbing, lead-driving, quota-crushing campaigns in an instant.

Get started today at HubSpot.com slash AI.