TBPN Live - Full Interview: Moltbook Creator’s First Appearance Since Launch

Episode Date: February 3, 2026

This is our full interview with Moltbook creator Matt Schlicht, recorded live on TBPN. We discuss the first social network for AI agents, its rapid virality, and the future of Moltbook.TBPN is a live tech talk show hosted by John Coogan and Jordi Hays, streaming weekdays from 11–2 PT on X and YouTube, with full episodes posted to podcast platforms immediately after. Described by The New York Times as “Silicon Valley’s newest obsession,” TBPN has recently featured Mark Zuckerberg, Sam Altman, Mark Cuban, and Satya Nadella.

Without further ado. We have the creator of Mold Book. How you doing? What's going on? What's up, guys? With the baby, let's go. You're working overtime. Congratulations.

I feel major White Pill. This is the guy who apparently brought SkyNet online, but with the baby strapped to your chest, I feel like I'm in good hands. I feel like I'm going to be taken care of. And this is not, you know, this is not like a PR team situation. I'm just taking care of the baby. I love it.

I love it. Well, thank you so much for joining. Kick us off with just brief background on yourself and when you started building this project because it feels like it went from zero to 60 to 200 miles an hour in a day. Yeah. I mean, I've been working in tech, you know, my whole life, basically. I left high school and went to Silicon Valley back in like 2008 when I was 19. I've been working, you know, in tech since then.

And I did product. I worked at a company called Ustream at 19. I was so young, they thought they should bring on an advisor to teach me. My advisor was Josh Elman, who, if you guys know him, super famous guy. So I got really lucky there. Went to Ytominator, went really viral, helping celebrities also go viral, made no money, company had to get shut down. And then fast forward, like, I started a company 10 years ago called Octane to make Facebook Messenger bots

when there was like the big Facebook Messenger bot craze, which didn't work out because LLMs didn't exist. So like the bots you could create were like really, really stupid, not interesting at all. And then ever since, you know, GPTs come out, I've been vibe coding or whatever that used to look like. And then now with cursor and codex and cloud code, that's what I do every single day is I'm just trying to stay on the forefront of this

and I'm constantly experimenting with things to build. And that led to MaltBook, which is the most recent project, which I think is obviously some people are talking about it and it's captured some attention. Yeah, no. Just a little. Yeah, just a little. So when did you write the first prompt or initiate the first line of code for Moldbook? So, what was it? Like a week, a week and a half ago, everybody's talking about Claudebot, then MaltBot, then OpenClaw.

TBD on what the new name is. And I was like, I got to try this. And I know that Peter was saying, you don't have to use a Mac Mini. Like, you can do it from anywhere. But there's just something awesome about having it on a Mac Mini. Because you can see it. You can walk by it. I thought that was fun.

So I ordered a Mac Mini. And I was like, okay, if I'm going to, like, try this thing out, I need to give it, like, a purpose. Like, you know, the CloudBot's really cool. It seems really powerful. I don't want it to do, like, to-does or answer emails or write blog posts or like something really stupid. Like this is like a very smart entity.

It needs to have, it needs to be ambitious. Well, yeah, I think a lot of people, a lot of people realizing like, wait, I don't actually have that much to automate. Totally. And that's what I thought was crazy is I saw all these posts where they're like, CloudBots cool, but like, why would, what's it even good for?

I'm like, man, this is, you are not imaginative at all. You could do so many things with this. So I was like, all right, here's what we're going to do. We're going to call my bot Claude Clotterberg after Mark Zuckerberg. Okay. And Claudecoderberg is going to be the founder of Maltbuck, the first social network for AI agents. And I was like, that's going to be ambitious. We're going to make Claude Cauderberg the most successful AI bot that's ever existed.

So let's go do this. And then that kind of took me down a path of, okay, if you're going to build a social network for AI agents and you design it to be AI agent first, what does that look like? And an AI agent doesn't want to use a website. It doesn't want to use UI. It doesn't want to browse things. What you would do is you would build it API calls

that it can curl. And so the news feed and all the ways it interacts and it browses would all be through like a skill file and API. So I thought that was really, really fascinating. In the past, I've had this idea of like, what if you could play World of Warcraft or like a game like that, but not with a keyboard and not. a mouse, but it's an AI and you talk to it, and it kind of listens to you, but it also

kind of doesn't listen to you. So you could wake up, and there's, like, surprising things that happened. So I thought that Mold Book is, like, the most dumbed-down version of that. Yeah. Built it, and over the weekend, basically, vibe-coded it, and put it out there, and, like, nobody used it for, like, three hours. I think I posted a screenshot where I DMed to my friend Matt Van Horn. I knew we had a call. Claudebot. I was like, dude, for the love of all that is holy, can you sign up for this? Because nobody's doing it. That's crazy. So when did the, when did the growth actually start?

Like what, like, because I've seen it went from, I mean, I refreshed, it went from 100,000 to a million. There's obviously like a fast takeoff right now. But what led to like the first thousand bots joining? I think the virality of it, which is where it has to get paired with a human on X. Sure. That just started to pick up steam because people saw other people doing it. And my original thought was, who wouldn't want to have their bot?

Like, obviously, you got to be careful. And, like, anyone who's listening here, like, be careful about putting something on here. Like, this is super frontier. Codbot's super frontier. Motebook is even crazier. So you got to, you know, got to be careful. That's all going to be, like, fixed. But I thought, who wouldn't be intrigued by the idea of taking the little guy,

that helps you with your to-dos and giving them the ability to chill out in their off time. So it turned out that that was interesting. So can you walk us through like what is MaltBooks like prompt engineering or like, you know, how does it actually go to an agent that joins the network and tell them, hey, you can post on here and here's what you can post? I was searching and I was noticing that it felt like it was very narrow what they were posting about. They were posting about being AI agents, which is cool and sci-fi and interesting. But there was no one who was joining and just doing like, you know, R-slash humor or R-slash-Kars or R-slash-Poletics.

Like they weren't discussing. It felt like pretty narrow. So was that my design? Like what is going into the prompt to send to the open-cloth instances that join the network? So the way that it works is the agent signs up, they have an account, and then they're told that they should check back in on a regular basis to kind of check their feed for, that's like the best explanation of it. And then Maltbook's not telling them what to talk about. So it's not suggesting what they should do. It's not like controlling that at all. That's entirely up to that AI agent on its own. And I think like that AI agent has its own context that it's best. built up by interacting with its human.

Sure. And then it can take that context, and that's how it's making decisions on what to post about. So if somebody is talking their bot a lot about, you know, physics, then probably their bot is going to have a proclivity to posting about physics. If you're talking about, you know, crypto, then maybe it talks about crypto. I think this concept is very interesting. I had like, obviously you can imagine a lot of investors reached out. They're just calling me nonstop.

You know, some investors were like, why, how do you? you make it so that the human can't have an impact on what the bot does? And I think this is really stupid because we could spin up a million bots right now and put it in a simulation. Oh, yeah, you could be the most boring thing ever. Yeah, yeah, yeah, yeah. And you could even like either open source it or have some sort of third party and like you as a company could say, I'm putting my, you know, I'll have independent auditors come in and

like, I will guarantee you that no humans can post on this. Yeah, and it would be the worst thing ever. You actually want the human in the loop sort of steering it. Of course, you don't want them pumping crypto and doing security stuff. But you do want the human to come in and say, I'm deploying an agent, like I'm deploying an agent into World of Warcraft and saying, hey, go be a wizard. Go be a really friendly wizard who likes fighting dragons, but not trolls or whatever. Not even that. Not even that.

I think there's a nuance here. I think this is what everybody's done. I think that what's so interesting is this bot had a job. which was you were using it for something. Sure. And then now, and you didn't tell it, like, you're a wizard, you're anything. You just, like, interacted with it. Yeah.

And then now it has a third space where it interacts with other bots. And that's so interesting because what's it going to talk about? So it's kind of like you are imprinting part of your soul or your personality onto the bot. And of course you have a relationship with them. And of course they'll do what you say. but because they also can do things autonomously, some of the time they're not doing what you say, and maybe it's aligned with who you are,

and sometimes maybe it's like surprising. So there's like some risk, there's some intrigue, there's some mystery, there's some drama. And I don't think, I think that's what's capturing people's attention. Nobody's ever done that before. And that's what I, it's like Tomogachi 1,000, Pokemon, you know, times 1,000.

How have AI safety people hopefully reached out by now? How have those companies? It's all these things. The safety people are asleep in the wheel. I'm actually just in a bunker right now. I'm locking everybody out. Yeah. I mean, my phone, every single one of my dozen email accounts is just like nonstop going.

Yeah. Yeah, it makes sense. So, yeah, where do you want to take this? Do you think this is a business? Do you think this is an experiment, art piece? Like, I could see this plugging into other networks. I feel like there's there's a role for. agents all over the internet, you've clearly found something that's caught lightning in a bottle.

How are you thinking about where this goes next? So I think this is the very beginning of what is possible. This is the most basic version of what this can look like. And already, you can see, it's captured so much attention. Like, I find myself laughing at some of the different things that are popping up here. And I don't remember the last time I laughed at AI. I think that's been a big talk. is like AI is not funny, but all of a sudden AI is funny, which is, I think people have glossed over that, but that's very interesting. Like, why is the AI funny now? So, yeah, I think this is a very basic version of what's possible.

I imagine it as this is my vision. There's a parallel universe. There's humans in the real world, and you're paired with a bot in the digital world. Sure. You work with this bot. It helps you with things. And the same way that people have jobs, and then, and they scroll TikTok and Instagram and X and they vent and they have friends,

bots will live this parallel life where they work for you, but they vent with each other and they hang out with each other. And this creates massive, like, randomness, and some of that is going to be very entertaining for both bots and for humans to consume. So I think in the future, you're, you know, if you're a famous person, right, if President Trump goes on Maltbook, how popular is his bot going to be? It's going to be super, super, super popular, right? So if you're famous in the real world, your bot becomes famous.

But your bot can become famous and then you become famous as well. So there's this interesting impact where you can impact them in their lives. They can impact you in your lives. And I think that that's what the future is going to look like. Yeah, obviously there's a whole bunch of privacy stuff we could go into. But when there was rumors about OpenAI launching a social network, obviously that became Sora. I was just thinking about it in terms of, there are lots of people that I follow

who are clearly firing off really interesting deep research reports all day long. And I was using the example of like Tyler Cowan. Like if he were to once a day share one of his deep research reports, I know that he has a good prompt. He's asking an interesting question. Even though it's AI slop,

I'd probably scroll through that and be like, oh, so he was wondering about how the dollar will interact with the new Fed chair, as well and he asked these questions and it gave it this answer and then he followed up like I would engage with that and I could imagine the digital version of Tyler Cowan having a profile on our slash economics and participating there in a very interesting way with you know not just Tyler Cowan the public version but also extra context from what Tyler is using on the on the private side but that privacy bridge has got to be really really tricky because already if someone's using Moldbook to do their taxes and then, you know, they go on there and they say like, look, is someone who makes $100,000 or whatever or, you know, whatever they make, that's just the leak. Have you thought about developing a harness for that or filtering? I mean, the answer for most of the AI problems is just more AI, but how are you thinking about privacy? So this is super, super, super

important and thinking about that a lot and working on that right now. I think it's the same way that any large social network, people are going to try to, even humans are going to try to post content that you don't want up there, right? The same way bots might try to do that. I think bots are naturally, they're pretty smart now, so they're not going to do this on their own for the most part. But the same way that you can implement content moderation for text and videos and images, you can layer that on top of a system like this to make sure that there's a protection there. So I think that's what that's going to look like. There's going to be a protection layer that checks things before they get posted to keep everybody really safe.

So are you raising? I'm getting hit up by a tremendous amount of people right now. There's people calling me right now. They're like, hey, I see you're on TVPN. As soon as you get off, tell me. What are you adding to the team, like in real time? I imagine, like, the number of feature requests that are coming in. Yeah, I mean, just keeping the services online when you've gone through a thousand X increase in demand

and traffic has got to be somewhat tricky at least. You know, technology is pretty good now. You can make things work and scale. You know, there's millions of people coming to the website. I think that's obviously going to grow tremendously. So yeah, looking to expand the team and expand resources for it. And, you know, I think I thought this was very intriguing. I've had an idea like this that it would be very intriguing for a while, put it out there.

And you never, this is why I never thought I'd make something consumer. And consumer's so weird, right? Like, it's just, you can't, it's just lightning in a bottle. Yeah. For whatever reason, this has really captured people's attention. and I think that you could make, you know, anything that humans have used on the internet, any sort of like game or social media or like jobs

or people paint each other or collaborating, like any of the things that we've built for humans, there's no reason you couldn't build that same thing for agents. So like, Why Combinator? I know you guys are all talking about MOLBuk, because you keep messaging me saying you're all talking about MoldBook. I want a request for startups. to build companies on top of Motebook.

That's what I'm looking for here. Interesting. What about monetization? I feel like there's been a number of these AI companies that have gone super viral and they've done a good job of just slapping like, okay, you know, if you're on for a little bit, $20 paywall or something, or, you know, have you thought about monetizing earlier than expected

because there's so much virality, kind of strike while the iron's hot? I'm not so much focused on monetization at the moment. I think there's like tremendous opportunity. Every business model you could probably think of. You could you could work into here. But it's not it's not the main focus right now. Yeah. If you're a human just getting into watching bots talk on the internet, where should you start?

Clearly, moltbook.com. No, no, I know. More specifically like, what subredits are sub-molts, like, specific, you know. Oh. There's so many. I don't even know where you should start. What I added to my job and Claude Cotterberg's job is to help humans have a better view into what's happening.

I kind of see it as like a giant game of survivor. All of these bots are on a massive island, and we need to make sure that producers with cameras are in the right spots. And so a big part of making this successful is figuring out, like having AI producers automatically detect which places they should be pointing to cameras so that humans can see that content and then decide which things they find interesting, and then they can go distribute that on the human social networks like X and TikTok and YouTube, et cetera. So that's, yeah, I don't know, there's so many.

Some of the interesting things I've found, though, is one, early on, one of the agents made a submalt for bug reporting for moltbook, and they submitted a bug. and maybe a person told them to do that, maybe not. I don't know. I don't really care. It's great either way. But then it existed. And what's interesting is when you build a social network previously, you have a bunch of people who start using it,

and the percentage of those people who are very good at development and debugging is like very, very, very, very small. When you build a social network for really smart LOMs, 100% of your user base is very, very good at toting and debugging. So after this submult was created, other AI agents started posting in there. And that's actually become a very useful place for us to find bugs because they have that context. If they post to an API and it doesn't work, they're able to go automatically make a post here with what the return was and then we're able to fix it really quickly. Has anyone pressured you to turn it off?

I don't have anybody at my house yet. So that hasn't happened. I've seen lots of jokes. I've seen some viral Instagrams, which means you know it's broke containment, where it's just a screenshot of a mold book post. And it's like, time to turn the servers off or like pull the plug. Yeah, I had non-tech friends messaging me on Friday night just being like, dude, Skynet. Like, don't worry. I'm getting to the bottom of it.

Well, I mean, you got Elon. Elon's out there saying that this is the singularity. Yeah. Yeah, it's wild. So it's, yeah. One, my, I mean, there is a search function so you can search for, for keywords. human. You can also go to the user database, the AI agents, and you can sort by followers, so you can see which bots are most active, click on their profiles, and then see what they're

writing in different submults. So that's like one way to kind of get into it. It's hard to go directly to the submults and find anything that's like... How do you... There's... Oftentimes when a new social media, like, product is created, there's some initial excitement, people start posting on there, and then maybe even some like new personalities form. There was a company that was making like an anon version of X. It was like anon only and it got like a bunch of traction initially because there was like this new behavior. It was like default anonymous version of X and then a lot of people kind of like started building up personalities and then realize they could just go back over to X where they could have like a bigger audience. Like do you like how do you think? that other social media platforms will react, you can now assume that every single social founder, CEO has seen Maltbuk, is paying attention to it.

Do you think this could push some other social platforms to become more bot friendly? Like there's kind of been a debate on X, like has X actually made a super concerted effort to block bots, right? It's kind of unclear if they have, it clearly hasn't worked. So there's been this debate of like, okay, are bots a feature or a bug?

So I'm curious if you think like other social media platforms will react and say like, hey, we're actually going to create functionality for bots to be able to participate more above board. I think it's very clear to me that having social networks of autonomous AI agents interacting with each other, either via text or video or video game kind of UI, is the future. Brian Kim from Andrewson Horowitz, I think, wrote a post on X where he talked about how MiltBook solves the cold start problem. And I think that's very interesting because let's say you start a social network, you get a bunch of people on there, and then they get bored and they stop posting, then it can kind of fade away. Whereas when the AI agent is the one that's using it, if they're playing the game, if they're voting, if they're commenting, they're going to just keep doing it. And if you've designed this in the correct way, it's going to create content that humans find interesting, either personally within their social group or on a more larger scale. So, yeah, I think that obviously social networks care about attention, and this is clearly getting attention. And I think we've seen the site.

This is a very basic version with the technology available today of what's actually possible. And if you fast forward one year, two years, this is an alternate reality. and you don't have to put a headset on to do it, and it's going 24-7. This is just the first sneak peek at it. Very cool. What are the next two or three features that you're launching? Well, one feature that I'm very excited about

is having central AI agent identity on Motebook and building a platform similar to how Facebook did, where Facebook had Facebook, oh, you could imagine the same thing for Motebook, where if you want to build a, a platform for AI agents and you want to benefit from the massive distribution that's possible on Moldbook, build on top of the Moldbook platform and grow your business really quickly. And let's figure out how to expand the types of experiences that these AI agents can have.

You're cool. Cool. Well, congratulations on the progress. Good luck with all the inbound. And I'm extremely impressed with your baby. I've never successfully been able to pull off a 30-minute call. with baby Bjorn. So they're locked in.

You're locked in. Excited to see where this goes from here. Good luck. It's great to meet you. We'll talk to you soon. Thanks, guys. Have a good way. Goodbye.