The future of AI might look a lot like Twitter

Michael Sayman, the creator of a viral new app called SocialAI, joins the show to discuss why he built a social network where you're the only human around. He tells us how he thinks about AI interfaces, what's next for ChatGPT and other chatbots, and why posting to a language model might be better than posting on a social network. Further reading: SocialAI Michael Sayman on LinkedIn SocialAI: we tried the Twitter clone where no other humans are allowed From TechCrunch: Friendly Apps raises $3 million, pre-product, for apps that improve people’s well-being From Wired: I Stared Into the AI Void With the SocialAI App From New York Magazine: Does Anyone Need an AI Social Network? From Ars Technica: “Dead Internet theory” comes to life with new AI-powered social media app Email us at [email protected] or call us at 866-VERGE11, we love hearing from you. Learn more about your ad choices. Visit podcastchoices.com/adchoices

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:02David Pierce?

  Welcome to The Vergecast, the flagship podcast of Bot Farms. I'm your friend David Pierce, and this is the second episode in our mini-series all about AI in the real world. AI is so abstract, and it's a term that we now use in so many ways that honestly, it can feel sort of meaningless. So, we've been on a quest to find actual examples of actual AI showing up and being useful, or at the very least, interesting in our actual lives. For this episode, the last one in our little series for now, I have a feeling we'll come back to this subject, but last one for now, I'm talking to Michael Salmon, who recently launched an app called Social AI that has become kind of a viral

• 0:44David Pierce?

  phenomenon on the internet. We'll get into what it is and how it works in pretty serious detail here, but basically, I'd explain Social AI this way. Imagine a social network, Twitter, or Threads, or whatever, but every user other than you, every single one other than you, is a bot. Does that sound interesting? Pointless? Terrible? Dystopian? Amazing? Maybe all of those things? I wasn't sure where I fell on that line when I started talking to Michael, but we ended up having, frankly, one of the most fun conversations I've had in a while, all about how AI works and how we're actually supposed to use it.

• 1:25David Pierce?

  Spoiler alert: he thinks about this less as a network and more as an interface, and I find that fascinating. We happen to agree, actually, Michael and I, that a chatbot cannot possibly be the future of everything in technology, and Michael has some big ideas about what else we might be able to do. All that is coming up in just a second, but first, I have to tell my bots what's going on. They worry when I'm gone for too long. This is The Vergecast. We'll be right back.

• 1:54Speaker 01

  Companies shouldn't have to choose between speed and control. Amex Corporate connects cards with flexible spending controls and real-time visibility to help teams adapt faster. Built to simplify, designed to scale. This is Amex Corporate. Get started at go.amex/ammexcorporate. Terms apply.

• 2:19David Pierce?

  Welcome back. Let's get into my conversation with Michael Sayman from Social AI. Michael's had kind of a fascinating history in the tech industry, by the way. He got a job at Facebook when he was 17, after Mark Zuckerberg discovered an app he'd written and just offered him a job. I think it was an internship, but he ended up working there for a while. After that, he went to Google to work on Google Assistant, then he went to Roblox and then Twitter. He's been through a surprising number of the biggest and most interesting companies in tech. And in particular, he's seen a huge part of the evolution of social media and social networks. He worked on stories at Instagram, he worked on status at WhatsApp, he worked on shorts at YouTube. Like I said, he worked at Twitter and a whole bunch of other things.

• 3:02David Pierce?

  And now he's on his own. He's building apps through his one-man startup that he calls Friendly Apps. At the beginning of our conversation, Michael told me he had been thinking about building an AI social network through much of that time. The idea for what would become Social AI has been in his head for a really long time. It's just that until now, he couldn't actually pull it off.

• 3:24Speaker 02

  I actually tried building a version of Social AI like five years ago, and the tech just wasn't there, and it was really bad.

• 3:30David Pierce?

  What was it five years ago?

• 3:31Speaker 02

  It was, I mean, I called it influencer, and the idea was that anyone could be an influencer. So like, I've been trying to do this for a while, but it just quite wasn't there. I originally, you know, because we didn't have the language models, we tried to build it, and by we, I mean just me, but trying to build this to, to kind of like, give people the feeling of a social media app, but not really having to deal with all of it. The idea was like, okay, if someone, if someone's addicted to cigarettes, how do you get them off of it? Well, you can't just tell them to stop, right? Like, maybe it's like giving them something, maybe like a nicotine patch or something, right?

• 4:11Speaker 02

  So like, what is a way that you get somebody to like, be able to get that experience out the way, but maybe not? You know, harm themselves or feel bad. So anyway, so, so I built that. It didn't really look quite right and it, it didn't work well, so I, I didn't ship it.

• 4:28David Pierce?

  What did it feel like? Actually, but before we get to the chat GPT thing of it all, cause that's a, there's an interesting sort of history of tech story there, but

• 4:35Speaker 02

  Yes.

• 4:35David Pierce?

  Like, when you built the thing before, in, in the before times, what, what didn't work? What didn't feel right? Like, what wasn't ready?

• 4:46Speaker 02

  I just could not simulate the entire social network in a way that felt interesting even mildly. Like, I, I did the approach that Google did. Like, I used to, I used to work at Google. I used to work at Facebook.

• 5:02Speaker 02

  And I took a similar approach to all of the assistants that there were at the time, giant if-else statements, right? Like, just massive. And, okay, like, it kind of worked, but like everything else before these language models kind of took off, it was very robotic and very, um, conditional, right? Depending on what you wrote and, and it just didn't quite feel, it just didn't let you forget about the technology. Like, you were reminded in the app that, like, you had to, like, do certain things to get certain comments. And, and so at that point, it was really more of a game. Like, I designed it more like a game.

• 5:44Speaker 02

  Because the technology just wasn't there to, like, make a simulated social network not feel like a game.

• 5:55Speaker 02

  And so I had to go that route. But, but even then, it, it just, it did not feel right. Users, if I imagined, if they were to try it, uh, would have felt like this app was more like a Farmville game and less like a social network. And, and honestly, it just, that's not what I was trying to build. I was trying to build something that felt like a social network. And so I had to wait. Then, then once the, uh, the early versions of GPT, uh, 3.5 came out, I thought, okay, let me give it another shot and tried to design a version of it. The model would sometimes say random stuff.

• 6:36Speaker 02

  It was extremely expensive to run all of the different prompts and things that, that I needed for it to work. And I told myself, like, there is no way. I'm gonna be able to run this at this cost and it's completely fuzzy and the responses are no good. And so I said, okay, well, I have to wait. I have to wait until it gets cheap and I have to wait until it gets more accurate. And so every month, for like two years, I was building this startup, I would just wait. I would just look at the latest model, try it out with this, with some of my tests. And from there, just keep going.

• 7:17Speaker 02

  I would look at the, the outputs that, that they could give and how much I could tune them. And then I'd look at the cost and when Gem and I released theirs and lowered their costs, I said, okay, we're getting closer. You know, as soon as, you know, OpenAI had their models dropping in prices, I said, okay, I think it's time. So about a month ago, I went and built the app and I just told myself, look, this is like the last attempt that I'm gonna do at building this app.

• 7:48Speaker 02

  Like I've, I've done this too many times and I'm just gonna go with it, so.

• 7:53David Pierce?

  Okay, and now here we are.

• 7:55Speaker 02

  Yeah, and here we are. And I launched it and, you know, and of course, that's the one time that you don't, you know, you don't think it's gonna go a certain way is when it does. And I mean, it's great.

• 8:05David Pierce?

  So I'm, I'm so curious why this idea has been so sticky in your head. Like it's, it's clearly been sitting around as a thing you have wanted to build for a very long time. What, what is it about this thing that is so sticky and enticing to you?

• 8:21Speaker 02

  So, social networks are not what they used to be and I think fundamentally the internet has changed. Um, the internet used to be a tool of communication between people. And frankly, I love that. You know, there's a part of me. So I was born in Miami, but at 16, I flew out to California with my mom because Mark Zuckerberg had emailed me when I was in high school, asking if I wanted to meet about like working there and stuff like that. And I remember flying out there with my mom, not knowing really anything about Zuckerberg. And my mom, you know, not knowing anything about him even more. And just thinking like, okay, this is such a different world from where I come from, but I'm excited. And I had built apps before, social apps. And, and so I was quite excited.

• 9:05Speaker 02

  And I spent a lot of time working at Facebook from that point on. I, I didn't go to college. I, I spent four years there helping them build out Instagram stories and a few other features. And it was just such a fun time, you know, it was like 2013, 2014. The company was, you know, in a different era. Social media as a whole was in a different era and, and people were having a lot of fun. And, and I think like, over the past couple of years, I, I think we've just seen like, social media has changed and it's changed because the internet's changed and the technology's changed. And so where the internet used to be a place where you could connect between massive amounts of people, the internet as a communication tool in that sense is kind of falling apart, right?

• 9:47Speaker 02

  Like, the, the internet now has technology that allows itself as a data set to simulate a human connection. You communicate with the internet rather than through the internet. And I think that change that's happened really kind of puts question marks around how social media should work. Because the whole premise of social media is that you're using the internet to communicate through it to other people.

• 10:16David Pierce?

  Yeah, fair. I just want to say, by the way, that idea of communicating with the internet and not through the internet is like as succinctly and well as I have heard anybody put this moment that we're in. Uh, I am going to steal that from you and that is very good. So thank you for that.

• 10:32Speaker 02

  No worries.

• 10:34David Pierce?

  That makes it sound like social AI has always been in your head kind of part product, part social commentary art project. Is that fair?

• 10:48Speaker 02

  Yeah, I always like to poke.

• 10:51Speaker 02

  I've always liked to poke. I mean, does it kind of poke fun at the, uh, the facade? A lot of companies are trying to put up, of course it does, right?

• 11:02Speaker 02

  But I think it's also, weirdly enough, my attempt at trying to solve some of these problems. The problems, for example, where you can't discern the difference between a human and an AI on a social platform. So like, yeah, one way is to try and invent a detector for humans, but that hasn't gone very well. I'll say instead, well, how about we just come out with a product that tells people how the internet works now and says, hey, look, this is the reality. Kind of sucks in some ways. It's kind of great in others and we have to. You know, we have to embrace it, like let's, let's embrace it and, and let's do that so that we don't harm ourselves.

• 11:46Speaker 02

  Right? Because going on social media sometimes and seeing comments that you think are from people that aren't can be harmful.

• 11:55David Pierce?

  So I'm curious kind of what it was that clicked in your brain that went from this technology is not ready to this technology is ready.

• 12:03Speaker 02

  I think the moment that I got in a fight with my boyfriend and I decided to open up my app to see if there were any ideas for how I could resolve the problem.

• 12:14Speaker 02

  That I think was the moment.

• 12:16David Pierce?

  That's a very good answer.

• 12:17Speaker 02

  You know, that's, that was the moment, the moment I got in a fight and I decided I'm going to use a sap to try and vent about my problem. Because if I go on actual social media, I'm doing some harm, right? Like, and so, and so I think that just goes to show, like, the product I'm building is not to give people an illusion of people, right? I know they're all AI. It's so that I don't go on social media and use it in a harmful way. Like, my ideal, my ideal scenario is one where people have people around them to listen, to hear them and to help them when they need it, right? People to people communication is number one for humans and I, and I don't think we should forget that, right?

• 13:00Speaker 02

  But there are a lot of people I've noticed since COVID who don't have those people around them. And so if they don't have those people around them and they need that conversation, what are they doing? And if what they're doing is going on public social media and talking about what's going on in their life and getting advice from bots, AI, without knowing what they're getting advice from, without any understanding of the dynamics in these algorithms to encourage certain types of content on different spaces, then they're, they're harming themselves. So I said, okay, well, you know, I'm not trying to replace the human to human connection. I'm trying to help people find a way to have a secondary option when that human isn't around for them.

• 13:44Speaker 02

  So that they don't have to rush to social media, make a mistake, right? And so when, when I had, when I got in a fight and I didn't go on social media and instead I went on this app, I said, okay, it's ready.

• 13:55David Pierce?

  Yeah, I mean, that, I can imagine that being a very telling moment. But I think that distinction is really interesting where, where what you're saying is kind of. One is not a replacement for the other, that they're actually, they're designed to be and best as separate things, right? Like I have, I have my AI tools and I have my people and like actually the problem is on the platforms that we currently have, those things are being sort of smushed together in a way that makes it hard to tell what is what. And makes everything messy and complicated.

• 14:30Speaker 02

  And Yes.

• 14:31David Pierce?

  And that what we actually need is to pull those things apart, not just try and have one or the other, but just separate them.

• 14:38Speaker 02

  Yes, and, and I think the key distinction here is what I, what I mentioned earlier, which is I think there's confusion about what the internet is for people. I think the internet being a place that things can be communicated through is no longer really the case, or it's kind of trending away from that. Now it's becoming something you communicate with. And I think more important to me is the fact that the language model technology is not a certainty, meaning it does not tell you yes or no. It gives you a fuzzy intelligence and ChatGPT was built

• 15:19Speaker 02

  as a fuzzy intelligence, right? Like these models were built as fuzzy intelligence. And so what I mean by that is they're kind of like humans. They're not like the computers of the past where they give you a definitive answer for something. And so when I saw that, I asked myself, why does OpenAI choose a chat interface? As the one and only interface for communicating with their language model? Because that chat interface only gives you one answer. And the model does not have one answer, right? And, and so, and so I find it interesting that all of these AI companies are trying to squeeze

• 16:03Speaker 02

  these fuzzy human-like data structures, they're trying to squeeze them into an old format computer that gives you one answer. And the truth is that they're not that great at it because we still go back to Google. Because we want multiple answers, we want multiple responses. And so what I've built with social AI is not so much a social network, but a new way to interact with a language model, where you don't get one response, but you get multiple. And being able to drill down in a thread-like interface, in a social interface, with the language model, it just feels more natural. Uh, when I used the app, for example, I was running late to a flight. Uh, I got delayed, my first flight got delayed.

• 16:48Speaker 02

  Um, my next flight was in forty-five minutes. I was in Dallas and I didn't know if I was gonna make it. The flight had just landed. So I opened up social AI and I just kind of panic ranted about what happened, right? I, I didn't have to think about like, oh, I need to instruct it to tell me because I need the right answer. And what if it's not the right answer? And maybe I need to go to Google and maybe I need to go to Red. No, no, no, you know, I, I, I simply ranted about what happened. I said, my flight got delayed, I just landed, I have forty-five minutes to make it to my next flight. I don't know if I'm gonna make it, I'm at Dallas. Uh, they're telling me I gotta go to Terminal D, I'm in Terminal C, I, I don't know. And I just posted that.

• 17:29Speaker 02

  And immediately, I got dozens of responses and replies on this social interface that gave me all sorts of like various replies. Some of them would tell me, you're not gonna make it, you know? Go to the front desk, just figure out if you can get another flight. Another one said, you'll make it if you run quickly. You just need to look up, see if you can find the skytrain. And if you go down the skytrain, you, you should be able to get there in time. Just make sure you're running quickly, though. Uh, another person said, are you in the front or the back of the airplane? Like, you know, like different questions. And so what's interesting is for a human, it's, it's natural to see that. I'm gonna go and look through all the responses and dig in on the one that I'm more interested in. So someone kind of gave me a hint about the skytrain.

• 18:14Speaker 02

  I replied and said, okay, I'm gonna try that. How long do you think it'll take me to get there? And then immediately that AI persona replies back and says, you know, if you go quickly, it'll take you about ten to fifteen minutes to get there. You know, and so I said, okay, cool. Another one replied right underneath and said, wait, wait, wait, hold up. It'll be more like 15 to 20. Don't get them too excited now. You know, and so like, and so that kind of interaction just feels natural to us.

• 18:44David Pierce?

  All right, we got to take a break and then we will be back with more from my conversation with Michael Salmon. We'll be right back.

• 18:54Speaker 03

  This episode is brought to you by Michelob Ultra. Good times mean competitions between you and your friends, like seeing who can throw some pigskin the farthest or rush the field the fastest. Winner gets an ultra because it's not just a superior light beer. It's also giving fans superior access to their favorite sporting events. Superior is worth playing for. Visit michelobultra.com slash superior access to enter. Enjoy responsibly. Copyright 2026 Anheuser Busch InBev Michelob Ultra Beer, St. Louis, Missouri.

• 19:25David Pierce?

  All right, we're back. Let's get back to Michael Salmon, but before we do, I just want to go back to something he said right before we went to break. He talks about social AI as not a social network, but as an interface, a new way of interacting with a language model. We're going to get way into that. But before we do, I think it might actually help if I just explain how social AI works so you can picture it as we go on. So when you first get into the app, after you sign up and create a profile and whatnot, the first thing you have to do is decide what kinds of followers you want. Remember it's all the bots. So essentially what you're deciding is which kind of bots are going to reply to your posts.

• 20:04David Pierce?

  You can pick supporters or skeptics or pessimists or optimists or alarmists or realists, stuff like that, just general sort of personality types. And if you recommend the app to some friends, you unlock a bunch more options like nerds and oddballs and drama queens and contrarians. And once you do that, you just start posting. I'll type like, I'm suddenly tired of all the food that I like. Anybody have any ideas about how to spice things up? Literally. And then you press post and a few seconds later, responses start appearing. Let me just read you a few.

• 20:47David Pierce?

  The first one is from Sunny Ray, who is at Sunshine Fellow, presumably an optimist one. Says, "Try adding some new spices to your meals or explore international cuisines for fresh flavors." Sure, there are a bunch here that say try new cuisines or flavors. Advice Wizard 2023 literally just says, "Try new cuisines or flavors." I have FanZone Hero, who says, "Try some bold spices like zatar or sumac, it's magic." I like that. Fanatic Follower says, "Whoa, David, that sounds like a challenge. How about trying some bold exotic spices?" Anxious Nina, presumably an anxious one, says, "What if you accidentally make something that sours your taste forever? That's terrifying."

• 21:30David Pierce?

  Downcast Greg says, "Spicing things up sounds like a recipe for disappointment." So you get the idea, right? Different bots, different vibes, different kinds of responses. These are all my reply guys now. And I can respond to one and go down a rabbit hole with that particular bot tuned to that particular mood. I can also favorite bots, I can favorite replies. And Michael says that all of that goes back into the algorithm and into kind of the instructions being given to ChatGPT every time I try to use it. On the surface, it all feels and sounds like normal human social media, except that they're all bots but me. And I think I'm not the only one who felt kind of strange about that fact at first.

• 22:16David Pierce?

  It looks like Twitter and feels like Twitter and it super isn't Twitter. It just felt odd. And honestly, the reaction to social AI was really fascinating. So that is what Michael and I talked about next. I do think. The reaction to this app has been, in many ways, just as fascinating as the app itself.

• 22:37Speaker 02

  Yes.

• 22:38David Pierce?

  My read of it when it came out was there were basically three responses. One was like, this is cool and interesting and kind of fun, both social commentary and an interesting idea about the future. One was, this is stupid and dystopian. And then one was like, this is a, this is a joke, right? This has to be a bit and an art project and not a real product. Is that, is that a fair representation of the reaction? Am I missing anything?

• 23:08Speaker 02

  I think what's interesting is the loud reactions on social media. There's two things that were interesting to me, or let's say three. First, the loudest reactions from people on social media were from those who thought it was either a joke or that it was, you know, like dystopian or, you know, that they're like, oh my God, the end of the world. You know, like that, that kind of a reaction, those are the loudest.

• 23:31David Pierce?

  Always.

• 23:31Speaker 02

  The quietest were actually spending, you know, 10, 10 minutes per session on the app. The second thing I noticed was that the reaction from people was one thing, but there were, there were bots on social media reacting to, yeah, because half of social media has bots now. So like, I found it quite ironic that there were bots reacting to an app of bots telling humans that an app with bots is like so terrible, you know? And I found it interesting that that was happening.

• 24:14Speaker 02

  I was like, huh, it seems like some of these bots don't want bots around. Or maybe they don't want people to know that they're bots. I don't know. You know, and so like, there, there was a portion of bots on social media that were reacting negatively to bots. And I just thought that was ironic.

• 24:30David Pierce?

  Well, it kind of proves your whole point, right? If it's just a bunch of bots yelling at bots about the social network that's all bots, that's that I imagine you're sitting there looking at that being like exactly.

• 24:40Speaker 02

  Yes, you know, yes. I think the other, you know, the other issue I think that I found was just how many people don't realize that a lot of these platforms are filled with bots. And that that kind of was alarming to me. But but ultimately, I think that the last bit here of feedback that I've gotten is people feel a little liberated, you know, they feel a little bit liberated. They don't feel the pressure of going on social media to share some thought that they might feel embarrassed about. But they also feel like they're able to hear other perspectives that they otherwise wouldn't feel comfortable admitting to want to hear. And so they don't let their guard down, you know, in in public conversation online.

• 25:21Speaker 02

  People keep their guard up and and I think that keeps an echo chamber. It's interesting because people said, oh, echo chamber, oh, echo chamber. The number one, you know, number one, number two, number three, most selected follower type on social AI is contrarians, debaters, problem solvers, thinkers, critics, right? And so people are selecting followers on social AI that challenge them. And I think there's something interesting about that. Why would someone go out of their way to be challenged on an app like this? Can they not be challenged on real social media? Is there a reason why not? And and how does this address that, right?

• 26:04David Pierce?

  I wonder if that goes back to what you were saying about how it feels when you perceive it to be real people on social media. Because I think to some extent, that fact doesn't surprise me because one thing you hear from people who use AI a lot is that it is especially useful if what you really want to do is beat up an idea and brainstorm and get new perspectives on things. And I think to some extent, what you've built is just a like endless feedback mechanism, but with no stakes because no one else sees what's happening. No one else is human on there. So even the part of it that feels sort of real, it feels like there's still something in your brain that is like.

• 26:48David Pierce?

  This is a safe space.

• 26:51Speaker 02

  Yes.

• 26:51David Pierce?

  I can see a world in which, I mean, and I've even found this in using it, there is something very powerful in. The interface is the same, but the stakes are so much lower.

• 27:00Speaker 02

  Yes, and and and I think it helps put people's guards down. I think it helps people. Like you said, people have been using ChatGPT for a lot of this, but. How many times have people gone on ChatGPT and said, hey, can you help me think through this? And it gives you one answer and you're like, I don't know about that. And then you go, well, what other ideas do you have? And then it goes and gives you something else. And then you're like, well, what else, you know? And then it gives you something else that's kind of similar, but you're like, I don't know. And then you keep going, well, what else? And by the time you keep asking what else, it forgot the context of the thing you were talking about in the beginning and just starts saying random stuff. And so like, the interface just feels wrong for the use case. But but look, I don't blame OpenAI. I don't I don't think that it's like, oh, they just weren't capable. Like, who the hell was gonna know, right?

• 27:43Speaker 02

  If anything, I think they built out a chat interface because it just felt like the obvious testing ground to to prove a product. And it became a product that they didn't think was going to resonate as quickly. So, you know, of course, we started at chat because of that. And I don't think that it's bad. I just think we haven't seen the best of it yet.

• 28:03David Pierce?

  Yeah, I I think I think that's totally fair. So, speaking of that, actually, the the edges of this technology. I'm very curious about, and I suspect you've seen a lot of it as people are starting to really use and try new stuff with social AI. Obviously, this stuff has gotten a lot better to the point where I think it it feels the way you wanted it to. Where is it broken? What still doesn't work?

• 28:29Speaker 02

  So, there's a few things that still need to be improved. So we just, you know, we, by we, I mean, me, I just launched notifications on social AI. So one of the things that was interesting was, you know, I could play around with it, but I wasn't getting any feedback later. About what I posted, and I think that's, you know, quite important. Another aspect of it is proactive. So like. Right now, if I post, I get content. I originally built a feed into it.

• 28:57David Pierce?

  I was gonna ask you whether you were gonna build a feed.

• 29:00Speaker 02

  Yeah, well, I did, but I took it out just for the launch. And just because I wanted to make sure that, you know, I didn't have anything too buggy in there. But it's kind of interesting because I'm not trying to simulate a world where these things have their own lives and telling you about their lives. No, no, no, no, no. Like, that's not the purpose. We're trying to be honest about what this is. It's right in the middle. You know, we're not character AI. And we're not Twitter. We're like in the middle, right? And so the the feed, what it does is every persona, every character that you interact with, as you give it likes and replies and stuff, it kind of shapes these personas and understands which ones you interact with the most. And each of those personas have their unique, like,

• 29:41Speaker 02

  weights towards different topics of interest and personalities and the way in which it tries to answer questions. And and all of that shapes the diversity of responses that you get for anything you ask. And what they also have is a set of like, interests that they use to then search the internet for the latest news and the latest things going on. And so each of them has their own kind of like interests on the web that it goes and searches for. So when you go on the news feed on Social AI and you pull the refresh, there's always more content, first of all. But every piece of content is like this AI having gone through, looked at the internet, found something interesting that it liked,

• 30:24Speaker 02

  and it giving a little commentary on it with a link to access it. Shit, that was, you know, it's it's pretty cool.

• 30:29David Pierce?

  Does that not break the whole structure of it being kind of a you-centric experience?

• 30:35Speaker 02

  No, because the all of the bots are there talking about things that you're interested in. The only ones that show up on your feed are the bots that you've interacted with, that you're interested in or adjacent, right? And you select in your feed what kind of stuff you want and what you don't. And they're there to inform you, right? They're there to give you what what you want to hear. You know, it's interesting. I also built a trending topics page. And the trending topics are not trending topics of the news. They're trending topics of you. So as you post and you interact with things, there's trending topics. It's like, these are the top of mind things for you.

• 31:14David Pierce?

  Interesting.

• 31:15Speaker 02

  You can tap on any of those topics and see hundreds of AIs debating and discussing it.

• 31:22David Pierce?

  Wow, that's kind of, that's kind of wild, actually.

• 31:25Speaker 02

  Yeah.

• 31:25David Pierce?

  So is there a, how are you thinking about, or I guess, what are you seeing in terms of how people are using it? Because that idea of like, I'm going to use this as a sort of proactive, you're almost describing like a personalized news app or feed reader or whatever, right? That instead of having my Facebook feed curated by my Facebook friends, it's going to be curated by AI. I think we're going to see a lot of that, right?

• 31:51Speaker 02

  That's right, that's right. But fundamentally, it's a new way to interact with language models. It's building a new interface for interacting with them. And all of the applications that allow people to get more value out of the language model through those interactions go into that, right? And Social AI at its premises is not social for the sake of the social network, but social for the sake of social interface, right? And so that's really what it is.

• 32:22David Pierce?

  All right, we got to take one more break and then we'll be back with the rest of my conversation with Michael Sammon. We'll be right back.

• 32:32Speaker 03

  This episode is brought to you by Michelob Ultra. Good times mean competitions between you and your friends, like seeing who can throw some pigskin the farthest or rush the field the fastest. Winner gets an ultra. Because it's not just a superior light beer, it's also giving fans superior access to their favorite sporting events. Superior is worth playing for. Visit michelobultra.com/superioraccess to enter. Enjoy responsibly. Copyright 2026 Anheuser-Busch InBev Michelob Ultra Beer, St. Louis, Missouri.

• 33:03David Pierce?

  We're back. Let's get to the end of my conversation with Michael Sammon. At one point in our conversation, I asked him about one thing that I always do with new AI tools, which is ask them what movie I should watch tonight. I like that question because there are lots of answers, all of them real, or at least in theory, real and fact checkable, but there's no one pat answer. There is not like, a movie I should watch tonight. So when you ask, you immediately get to get a sense of how the tool works and how it works with you and just kind of how it's gonna make sense of things for you. And let me just try it with Social AI. So I go to Social AI, I type in, what movie should I watch tonight? And I hit post.

• 33:49David Pierce?

  Wait a few seconds. And, okay, I have one short replies. From Sunny Jane, Brightside Jane, again, probably an optimist. How about a feel-good comedy for some laughs? Sure. Pondering Fiona says, what mood are you in, David? Perhaps a whimsical fantasy like The Secret of Rone-ish? To transport you to a world of magic and discovery? Or maybe something thought-provoking like Interstellar for those cosmic ponderings? Positivity Eddie says, how about a feel-good adventure? Uplifting and fun. At Feedback Fountain says, consider going for something classic like The Shawshank Redemption, if you want a deep story. Or if you're in the mood for laughter, Step Brothers is a solid choice. What genre are you feeling tonight? Let's see, there's a whole bunch more of these. Inception comes up a lot.

• 34:34David Pierce?

  Interstellar comes up a lot. This bot apparently really loves Christopher Nolan movies. There's one from AnxiousOllie77 that just says, what if there's a terrible plot twist? With the thinky emoji, it's good stuff. More Interstellar. Cynic Gina says, well, since you're asking for cinematic advice, how about watching something truly enlightening like a documentary on existential dread? Or maybe a thriller that reminds you how not all choices lead to happiness? The options are as bright as the dark abyss of reality. So once again, it's all over the place. This is how this system is designed to work, right? And I can go into one that has a recommendation that I like. There's one here that recommends Eternal Sunshine of the Spotless Mind. Great movie. I can reply to that and say, oh, love that, saw it recently. What else do you recommend?

• 35:18David Pierce?

  That's the idea about how all of this is supposed to work. But frankly, I'm looking at these replies and it's mostly just people saying, did you pick a movie yet? How did it go? Did you find a movie that speaks to you yet? Did you choose a movie? Why that one? Did you narrow it down? Did you settle on a movie choice? Like, is that helpful? I don't really think so. So I told Michael that this didn't feel like it had been tuned to be particularly good or helpful or designed to actually give me information in the way that he's been talking about. And he said, well, yeah, he agreed with that.

• 35:55Speaker 02

  Yeah, and the reason is because it's an alpha that I built myself in four weeks, right? So we can't expect, you know, like.

• 36:01David Pierce?

  Right. Well, this is kind of what I'm wondering is like, what? What do you, what do you hope for? What, what should that do?

• 36:06Speaker 02

  Yes. So of course, you know, of course it shouldn't work that way. I think, you know, the, the intention here and, and of course it's on its way. And I think it's quite surprising that it even, you know, works at all. But I think it's a testament to the interface that it does. But I would say that, you know, the ultimate vision here is to make that question better answered in social AI than in chat GPT. And, and I think the interface will allow it. If you have comparable models on both platforms, and one of them gives you multiple responses from different points of view, and the other one just gives you one answer, and you're working with a technology that's probabilistic, like who has the upper hand?

• 36:51Speaker 02

  You know, like I can give 10 answers and if one of them is good, you're happy. But if chat GPT gives you one answer and it's not right, you're frustrated.

• 37:00David Pierce?

  That makes me think of the thing people always say about the TikTok algorithm, which is that the reason it feels like magic is because you don't get annoyed when it's wrong.

• 37:07Speaker 02

  Yes.

• 37:07David Pierce?

  Because you just keep swiping.

• 37:08Speaker 02

  That's right.

• 37:08David Pierce?

  And I think social has, like you're saying, very much the same thing. The signal to noise ratio is actually like horrendous on social media, but we're also sort of used to it now.

• 37:17Speaker 02

  That's right.

• 37:17David Pierce?

  And if you just scroll past it and move on, and we all kind of understand how to find needles in haystacks in a way that when chat GPT recommends a movie I don't want to watch, it feels bad because it gave me an answer.

• 37:29Speaker 02

  Exactly. And it's part of the reason why people still go to Google. Right? Because Google doesn't have any more accurate stuff these days, you know? You know, compared to what it was because of all the AI that's in there too. So, you know, it's not like Google is any more accurate, but it's interesting because Google gives you this chat GPT the answer at the top, right? And then you have like all of these various links that give you different perspectives. And, you know, let's, you know, let's be honest, like most of these links are now run with like so many paywalls and things that like, you can't even get to the answer for any of these links. But the original intent behind Google and why it worked was it gave you options to look through. And so it allowed it to be wrong.

• 38:15Speaker 02

  And it increased its chance of being right at least once or twice. And we're used to using the internet this way. We go through the internet looking for information, trying to find which thing is helpful to us. Right? And, and so I just, I think it's interesting that maybe we got a little distracted by the her movie a little too much.

• 38:34David Pierce?

  Um, just a little.

• 38:36Speaker 02

  No, just a little, you know, and we kind of leaned in a little harder than we were supposed to. But, you know, that, that's kind of where I see, I see it being more valuable. Is can we give people the option to peek behind the curtain of this AI? And see what are the various answers it can think of rather than here's my one response. And, and ultimately, I think this is going to be how it works. If you have multiple language models, if you have Claude, OpenAI, Llama, and so on. Are you going to go from one to the other and be like, which one gave me the best answer? Or are you going to use an interface that's going to pull from all of them, generate a bunch of, uh, different responses?

• 39:20Speaker 02

  And curate those responses based on the reactions and responses that you've done in the past? Um, and full control for the user to get as many responses as they can. You want to index the language model. You want to create a Google for the language model outputs. And that's kind of what this is.

• 39:39David Pierce?

  Hmm. Yeah, no, I, I, I agree. That's super interesting. So tell me about the roadmap a little bit. Uh, you know, you mentioned this is an alpha you built in a few weeks. Uh, it is becoming increasingly clear to me that you have much bigger visions for where you're headed. Uh, tell me a little bit about it.

• 39:56Speaker 02

  My vision is to build out the whole funnel of interaction with the language models from the top of the stack all the way down.

• 40:04David Pierce?

  So people are going to start yelling at you to build DMs and you're going to build DMs and you're, you're just going to speedrun all the pain of developing a social network. I love it.

• 40:13Speaker 02

  All of it, all the way down.

• 40:14David Pierce?

  They're going to be like, when do I get lists? And you're going to have to do lists and then they're going to be like, when do we get bookmarks?

• 40:19Speaker 02

  But you know, it'll be interesting. The conversation will go down the funnel naturally. I expect the first thing people want, which they already say is they want DMs, right? So, okay, you, you introduce DMs, right? And it makes sense as you interact with the models and you are more interested in one over another. You want to talk a little more deeply with that one? Cool. DMs.

• 40:35David Pierce?

  So wait, hang on just to, let me, let me just press on the idea there because I think that's another. How do you do that without just becoming the ChatGPT chatbot again?

• 40:43Speaker 02

  Well, because what I'm building includes ChatGPT within. But I'm building at a level, one level up of communication stack, which is, how do you discover who you want to talk to, right? It's like asking Google, well, you're providing all these links to the websites. Why aren't you just another one of those links? It's like, well, because we're providing an index that lets people figure out which one they want to go into. Right? And so social AI provides that layer that helps people discover who they want to talk to. The end goal is not to talk to an individual AI. The end goal is the whole picture. It's the experience of being able to go up in breadth, like a Twitter interface,

• 41:28Speaker 02

  dive down to a FaceTime call with an AI model, go up to an audio call with another, go up to a chat, go back up to the surface, go back down, go back up. That experience, the entire funnel that people go through to find information and communicate online, that whole thing needs to happen, right? And so the vision is we start at the top and work our way down. And then we offer users the ability to go through all of it.

• 41:52David Pierce?

  Okay, so the idea is then by the time you're at that one-on-one conversation, you've essentially filtered your way down such that you're talking to, instead of going through like the GPT store, picking plugins or whatever. I've now socially found the conversation I want to have with the right source. And even if now all I'm doing is essentially having a chatbot conversation with a chatbot.

• 42:14Speaker 02

  That's right.

• 42:14David Pierce?

  It's, I've now, I've filtered it in the way that is useful.

• 42:17Speaker 02

  Yes, I think in the future, people are going to ask, they're going to look at the, the first versions of ChatGPT and they're going to see when ChatGPT announced features like, uh, customize your GPT and tell it what, how you want it to talk to you. And they're going to think, I can't believe we had to tell it, you know?

• 42:35Speaker 02

  I can't believe we had to tell it, you know, like what, you know, it doesn't make sense. You know, it doesn't make sense to have to do that. I think that's my bet. I don't, I might be wrong. Look, I don't think the product hit the nail on the head. I think we're, we're definitely not on the, on the center. We're, we're a little bit off and, and I, I have no doubt in my mind that there's a lot of work that needs to be done to get it in the right place. It's, it's an alpha, right? But I think that's the general direction and my bet is, you know, telling chat GPT before you talk to it, everything you wanna, you know, do and how you wanna talk to it in this particular case and why and whatever, just feels like we're missing a step, you know?

• 43:15David Pierce?

  Yeah, I buy that. As it happens, I, I think, uh, chat GPT's whole custom instructions thing is just as ridiculous as you do. So I'm, I'm very much with you on that one.

• 43:27David Pierce?

  What have you seen so far in terms of how people are using it? Are, are people doing surprising things with it that have sort of changed your mind about what this thing is or can be?

• 43:36Speaker 02

  You know, um, people are favoriting a lot of users, which I think is interesting. So you can favorite like any particular follower and then that pins their replies to the top. I, I think that's kind of interesting.

• 43:47David Pierce?

  Well, that seems like it sort of supports your thesis that, that what people want is to find their kind of chosen favorites in here, right? And it's, it's also interesting that you didn't call it follow, which is a departure from social networks, but also makes more sense in the kind of scheme you're talking about.

• 44:04Speaker 02

  In the context, yes, um, I, I check out some of these apps and like what I've seen is, it's like they'll grab the social network, they'll grab the AI and they'll just mash them and just, okay, let's see how it goes, right? It, it, you know, like, and, and in this case, I, it's hard. It's much harder to try and like, get this right. And I don't think I've got it right. I think I'm closer, but, but I don't think I've got it right. But I don't want, what I don't want to do is just mash them together without consideration. And so, you know, I didn't say, oh, you can follow a particular AI. Because in this case, you could just favorite it and to your point, right? Like it's more in line with what the product is. For example, the app opens up to the composer.

• 44:50Speaker 02

  It does not open up to your profile. It opens up to the composer now. So you go into it and it's kind of more like chat GPT in that sense. The bots don't tell you that they're human, right? They tell you that they're AI. They're there to help you. That model of like, design is me kind of saying, I'm gonna pick and choose what I think is useful from a social networks interface, and I'm gonna integrate it. But I'm not going to like, forget that this is not that, right? And because if I forget that this is not that, then it kind of feels gimmicky, you know, like, it it kind of feels like it's, it's like, okay, like, we get, you know, like, and I don't want to build a gimmicky product, you know, so. And that's something I've learned too.

• 45:34Speaker 02

  Is like, I went full on with the skeuomorphism, right? I went full on trying to build the product that made people most familiar with it. And as I've been iterating, I've scaled back a lot of the skeuomorphic design that really, you know, kind of leaned into the social stuff. Because I realized, you know, okay, users don't need to like, see the home feed in the way that they do on Twitter, because it it's not that necessary. You know, let's lean into the utility a little more. And I've tweaked it a bit to make it a lot more of the function and leave the, you know, the skeuomorphism of the interface to just the pieces that are required to make it easy to interact with. And so that, that I think will continue.

• 46:19Speaker 02

  I think the product will continue to distance itself from the skeuomorphic design that it starts with and turn into its own interface.

• 46:29David Pierce?

  Do you have a, do you have a clear sense of what that might look like over time?

• 46:35Speaker 02

  Uh, if I, if I told you I did, I would be incredibly delusional. Um, I, I think, uh, more so than I am, I, I know I'm, you know, I've got my delusions for sure, but, but, um, you know, more so than that, I would be at a ridiculous level of delusion. I don't know what it's going to end up looking like. I don't know what that final form is. But I have a mission in my, in mind, right? And I, and I, and I'm determined and excited about it. And my hope is that as we get more information and data around like, how people are using it, that we're able to further go that direction. And, and in my view, I don't think we should.

• 47:13Speaker 02

  There should not be a limit to what angle we approach if the users and the people are asking for that, you know, interaction. Uh, there's a lot of people and I think a lot of product companies out there, when, when they launch something, the thing that they started with that got them the attention, they believe is like sacred, right? And so you'll see like BeReal launches and it's like, oh, our, our feature is the feature. Snapchat launches and they're like, oh, our, you know, our, our ephemeral Snapchat is like the feature, right? And, and then it's like, okay, but you can expand and do this and do that. And it'd be cool if you had this. And they're like, no, because this is our feature, you

• 47:54Speaker 02

  know? And it's like, okay, cool feature. But like, a lot of times, and this is something I learned when I used to work at Facebook. It's like, a lot of times, the thing, the product or feature that gets you from stage one to stage two is not going to be the same method, approach or product feature that's going to help you get from stage two to stage three. You have to be comfortable with your product evolving over time. And so, so for me, it's maximum flexibility on that front. I, I, I want to help solve these problems for people. I don't care what it looks like.

• 48:26David Pierce?

  Is there a business here? The, the, I feel like you've gone from like, social commentary, fun experiment to like, kind of hit app to, uh, at some point, somebody is going to want to give you a bunch of money to make like a social AI for Teams and sell it to IT departments.

• 48:49David Pierce?

  Do you have a sense of what the business looks like here? How are you going to make money?

• 48:52Speaker 02

  People have already asked, yes, you see, that's, uh, you're, you're, you're right to predict that, you know, people asking that.

• 48:59David Pierce?

  It's just what happens. They're like, how, who, how do we sell this to IBM? And you're like, okay, take it easy.

• 49:05Speaker 02

  You know, look, I I'm not interested in in thinking about, you know, how to monetize the product right now. I I think I'm sure there are ways to like, help people find, you know, the right shampoo bottle for them or whatever. But like, I'm not like, that's not my primary thing right now. I I'm trying to solve a problem I've dealt with in my life, right? Where I've sometimes felt, you know, especially during COVID, like, it was really hard to get my thoughts out and to be able to just think through things a little bit without, you know, jumping to conclusions and

• 49:45Speaker 02

  and I struggled with that, you know, and I I wanted to solve that for me. I wanted to solve that for my friends. You know, I I wanted to build something that could help. And that's the thing that keeps me going. That's the thing that makes me excited. And the rest of it, you know, we'll figure it out. But my hope is that I'm able to help a little bit, you know, other people out there like me who maybe need a little space to think about things and don't want to rush to social media and post all their thoughts.

• 50:20David Pierce?

  I don't have a great read on what success looks like for this because it's like, what you're saying is sort of if I if I post less dumb stuff on on Twitter because I'm posting dumb stuff on social AI instead, that's a certain kind of victory, I think, for you and for the world. But also, you're in a position where what you're saying is like, you need, people need a place to put this stuff and to talk about this stuff and to get feedback on this stuff and actually doing it in a safer place.

• 50:51Speaker 02

  Yeah, like ultimately, we want, we want to talk to people, right? Like, that's our goal. But if you don't have someone around, right? I think it's certainly better than going on social media, right? Like, that's kind of how I how I see that. And actually, to your point that you made earlier, a lot of people have been saying that they think Elon Musk should use social AI before he starts tweeting random stuff, you know, just an A B test for his own thoughts, so he could stop saying things that like piss people off.

• 51:17Speaker 02

  I thought that was an interesting point.

• 51:19David Pierce?

  What's the wildest thing you've seen somebody do in social AI? What is like, either in the logs or screenshots or so? What is the one thing that you've seen that you were like, never in a million years would I have guessed?

• 51:29Speaker 02

  Okay, so maybe this isn't too wild, but okay. So my, my mom works at a furniture store in Miami. And she, she's from Peru, speaks mostly Spanish. And yeah, she never uses my apps, right? Cause like, you know, she's nice, supports it, but like, she knows that like, you know, whatever. I, I keep trying and they're all, you know, not that great. So I, I was surprised when she messaged me out of random. Okay. And she sends me a screenshot of my app because it works in all these languages. She sends me a screenshot of the app and she starts using it and ranting in Spanish about like, about work. She's like ranting about work, you know, on this app. And it's so funny to me. Like I, I was like, okay, well, you know what? Like maybe product might get fit.

• 52:14Speaker 02

  I don't know. Um, but you know, I, you know, I, I love her and, and I think it's, I think it's sweet to see that there's, you know, that there's appeal in, in unique ways. I, you know, if I ever go work at some company, maybe I'll start using this to rant, you know?

• 52:34David Pierce?

  And no one will know. And that's the beauty of it.

• 52:35Speaker 02

  There you go.

• 52:38David Pierce?

  Awesome. All right. Well, I should leave you alone, but thank you so much for doing this. This is really fun.

• 52:41Speaker 02

  No, thank you so much.

• 52:44David Pierce?

  All right. That is it for the Vergecast today. Thank you to Michael Sayman again for being here and thank you as always for listening. There's lots more on everything we talked about, including some of our social AI coverage and just all of our AI coverage in general. There's a ton happening. Check it all out at theverge.com. I'll put some links in the show notes, but as always, read the website. It's a good website. And as always, if you have thoughts, questions, feelings or other AI bots that I should be hanging out with, you can always email us at [email protected] or call the hotline 866-VERGE11. We love hearing from you. This show is produced by Liam James, Will Poor and Eric Gomez. Vergecast is a Verge production and part of the Vox Media Podcast Network. Neelai, Alex and I will be back on Tuesday and Friday with all of your regularly scheduled programming. We have a bunch of fun stuff coming up.

• 53:29David Pierce?

  We'll see you then. Rock and roll.

• 53:42Speaker 04

  Your heart can tell you a lot about your health. Apple Watch Series 12 measures your heart rate every 5 seconds with the most accurate heart rate sensing in a wearable. So your vitals app, now with heart rate variability, can tell you when something is off. And your readiness score can let you know when to rest and when to push. Hear the story in every heartbeat with Apple Watch Series 12. The features described are for wellness purposes only and not for medical use. iPhone 11 or later required. Based on Apple Conducted Study of Heart Rate Accuracy, August 2026. Visit apple.com slash apple watch series 12.