Full Interview: Clawdbot’s Peter Steinberger Makes First Public Appearance Since Launch

From Moltbox.

0:01

Man of the hour.

0:01

How are you doing? Thank you so much for staying up late. What time is it for you?

0:02

It's 11.

0:07

Thank you so much. We really appreciate it.

0:09

11Pm for everybody that's just here.

0:10

Yes, 11pm, 11pm so I'd love to kick it off with just a brief background on when you started this project, a little bit of your career, how you're thinking about it going forward, and then I have millions of questions.

0:12

This was your very first project ever, right?

0:23

Yeah, yeah, yeah. First time coding right.

0:25

Now. We were enjoying a screenshot of your GitHub profile earlier and just seeing how many different things.

0:29

An overnight success.

0:38

A true overnight success. But we're super excited to have you here.

0:39

Yeah, awesome. Yeah, I'm excited to be here as well. Yeah. I don't know. I worked for my own software company for 13 years and then I. I sold it about four years ago. Then I was completely burned out. I did like. I mean, it's. It's tv, but still it is blackjack and hookers.

0:42

Wild. Well, we're. God, we're glad you're back in the game.

1:12

Yeah.

1:15

Yeah. You know what they say, like for every four years in like one year break. And I did like 13 years non stop. So like three years, the math kind of checks out.

1:16

Okay.

1:24

And then this year. No, I mean last year, not 2016, in April. At some point, my spark was back.

1:25

Yeah.

1:35

Because before I was like, I was sitting on my computer and I don't know if you've seen Austin Powers, but it felt like someone sucked my mojo out. But yeah, I had time to recover. I came back in April and I wanted to do something new. My background was like a lot of Apple and iOS and I'm a little bit fed up. I wanted to build that stuff and I didn't have the experience. I didn't want to feel like an idiot. So I looked into AI and it was good. It was not great, but it was good. And I was like, why is nobody talking about it? You know, I feel like. Because I missed those three years where it was really bad. And I came back just at the time like Clock Code was released. What, February in beta.

1:38

Yeah.

2:25

So this was my first experience. I was like, this is. This is pretty awesome. And then I couldn't sleep anymore. Like, I literally had trouble. I had trouble going to bed. You know, we had like addiction before and then like we had addiction again.

2:25

But a positive one.

2:42

Yeah, well, yeah, I would say so. And I hooked up a lot of my friends for looking into AI as well. And they had the same problem. And I texted them at like 4am and they replied. I even started a meetup. That's where I come from. I called it Cloud Code Anonymous. Now it's called Agents Anonymous because you have to go with the times.

2:44

Sure.

3:05

And yeah, ever since then, that's what I say on my profile. I came back from retirement to mess with AI and I'm having loads of fun.

3:07

That's great.

3:17

I love it. Maybe walk us through some of the other stuff that you shipped and worked on prior to this and even just kind of like your mindset working on these different projects. I'm assuming at different points you would think that some would get more traction than others, but it would probably be impossible to have predicted in some ways that this would have gone from almost to the point. The reason that this is so wild is I'm seeing people on Instagram that like, I don't think of as people that like, follow tech at all. And they're at the Apple Store getting a Mac Mini. So it feels like it just went, it broke containment, like incredibly quickly. And you see the GitHub stars are like. Actually, I've never seen a chart like this every, you know, the last, the last, you know, every. Everybody loves to show their charts, but the chart is actually unbelievable. It's just a line going straight up.

3:17

I, I, I need to talk to someone at GitHub because I, I don't think that's been a project before that that's been like, straight. It is, it is, it is batshit insane.

4:08

Yeah.

4:18

I mean, honestly, my main mantra is I want to have fun. You know, like the best way to learn these new technologies, if, if you have fun with it, you have to play with it. So I build little things that I think could be useful. I tried different languages, I tried different approaches. It's gigantic engineering. I don't like the word wipe coding so much. I always make the joke. I do enchanting engineering and then when it starts hitting 3am, I switch to wipe coding and regrets.

4:20

You should have just gone to sleep, basically.

4:48

Sometimes that's hard. But then I just did little things. I had this idea about personal agents in May already and I tried. It was like the time that GPT4 1 was out, it was just not good enough. And then I thought, well, all the big companies will build this in the next few months anyhow. So it's like, why, why the F should I do that? You know, I was just going to wait and they make it better.

4:52

Yeah.

5:24

And I build a lot of stuff. There's, like, one project that is still unfinished that I. At some point, when I finish, and I build a lot of. A lot of cli. Because that's. That's where agents are really good. You know, you have to close the loop. That's always the secret you have to give. You have to build it so that the agent has the best possible way to build software. Yeah, this is. That's the. The secret a little bit. I tried a lot of stuff, and then in November, I looked, and still there was nothing. Like, where is. Where is my agent? I had a little project in May I spent two months on. Started as a joke because I did a hackathon with two friends, and we're like, what can we build? That could be kind of cool? Wouldn't it be cool if I could use cloud code from my phone? It's kind of like. It's something that everybody builds. I see this, like, every day. Like, by now, I almost call it like, this is like, one step in your journey in becoming a good, authentic engineer, is you're going to build some shitty orchestration tool for yourself because it's fun.

5:24

And you think, yeah, yeah, yeah, bridge.

6:35

And I built that for two months, and then I had to stop because it became so good that I was up with my friends, but literally with my phone, like, using cloud code to work on this thing. And it's like, this is bad for my mental health. It's already bad. And now I'm literally building something to have better access to my drugs.

6:37

Yeah. I mean, I saw. I've seen people using Claude code on laptops as they get off of airplanes because they're so locked in, they just have to send one more. And that's like, the clearest sign that, like, you need a bridge and a phone involved.

6:57

Yeah. No, but also, like, you know, like, this. This feeling when your agent's not running. Right now, there's, like, two terminals. He could be building something. Right. Yeah. So if you're in this addiction mode, you're almost like. You almost feel like if you need.

7:10

To step out for 10 seconds and fire.

7:22

Yeah. Feel free to take a break. We can do an ad read. Okay.

7:24

There's still some drama that I'm finishing, but anyhow, so in November, I don't know, I wake up every day, I'm like, okay, what do I want to work on now? What would be cool? And then they was like, okay, I want to chat with my computer on WhatsApp, because if my agents are not running and then I go to the kitchen, I want to check up on them or I want to do little prompts. So I just hacked together some WhatsApp integration that literally receives a message, calls cloud code, and then returns. What cloud code returns? One shot. It took, like, one hour, and it worked like, okay, that's kind of cool. But I usually use prompts, like a little text and an image, because images are like. They often give you so much context and you don't have to type so much. So. So I feel like this is, like, one of the hacks where you can prompt faster, just, like, make a screenshot. So the agents are really good at figuring out what you want. So I hacked together images, and then I was on a trip in Marrakesh with, like, a weekend birthday trip, and I found myself using this, like, way more than I thought, but not for programming. It's more like, hey, there's, like, restaurants. Because it had Google in it and it could figure out stuff. And it's like, especially when you're on the go, it is super useful. And I wasn't thinking. I was just sending it a voice message. But I didn't build that. There was no support for voice messages in there. So the reading indicator came, and I'm like, oh, I'm really curious what's happening now. And then after 10 seconds, my agent replied as if nothing happened. I'm like, how the F did you do that? It replied, yeah, you sent me a message, but there was only a link to a file. There's no file ending. So I looked at the file header. I found out that it's opus. So I used FFMPEG on your Mac to convert it to wave. And then I wanted to use Vispa but didn't have it installed, and there was an install error. But then I looked around and found the OpenAI key in your environment. So I sent it via curl to OpenAI, got the translation back, and then I unresponded. That was like, the moment where, like, wow. Yeah, you know, it's like, that's where it clicked. These things are like damn smart, resourceful beasts if you actually give them the power.

7:29

Sure.

10:01

And then I was kind of hooked. Like, I did all kinds of weird stuff. Like, I use this as alarm clock. I let it migrate to my computer in London, but then it used SSH to log into my MacBook and turn up the volume to wake me up in the morning. I think I built world's most expensive alarm clock.

10:04

Yeah, that's crazy.

10:23

And it even got it wrong because I had like. It uses a heartbeat. You know, like the concept of you do a prompt and you get something is already if this full access inherently dangerous. But I was like, let's turn it up a notch. Let's automate that. Let's give it a heartbeat. And the prompt was literally surprise me.

10:24

Wow.

10:46

But I see this project, as much technology as it is art and exploration, because this feels in one way it's just glue. It's just putting pieces together that we already have. In another way, it's. It's a whole different way how you interact with those things because all the technology blends away. You don't think about new session compaction, which model. I mean, maybe a little bit because tokens are still expensive. But usually all of that blends away. You just. You just talk to a friend or a ghost or.

10:50

Yeah. Last year everyone was wanting these agentic experiences. You were having this experience and it seemed like all the focus was on browsers and seeing the way that people have been using. Sorry, Multbots. Taking me a while to adapt. It just feels like all the focus was at the wrong layer. It's like, why do I care about the browser if I can just talk with an agent across every app, across every app, every surface. It's like I don't care about the browser at all anymore.

11:31

Yeah, I mean, a lot of the prep work I did before I built this was just build little CLIs. Because my premise is MCPs are crap. Doesn't really scale. People build all kinds of weird search things around it. But you know what scales CLIS agents know Unix. You can have like a thousand little programs on your computer. They just have to know the name. They call the help menu. They load in what's needed. We are calling the help menu. Then they know how to use it and then they can use it. And if you are smart, you build it in a way that just uses what the model already expects. You know, don't build it for humans, build it for models. So if they call minus minus log, you build minus minus log. As it's like agentic driven program. Like, yeah, build, build how they think and everything works better. It's a new kind of software in a way.

12:04

Yeah.

13:05

So for most of the things, I don't need a browser. Like I built something for the whole Google thing for places for my Sonos. I hooked up my cameras, my home automation system with every little cli and skill. My agent got more power and he got more fun. And I already had A lot of that working when I built the WhatsApp thing and I just got hooked. And the thing was, I find it amazing and I talked about it on Twitter and usually when I talk about projects I get response, but this one, it was very muted. It feels like people are not getting it. I showed it to my friends, even my non tech friends and they're like, they wanted it. So it was like I was up to something but the tech people wouldn't get it. So I tried a bunch of things. I kept working on it because I used it and ultimately I built it for me. You know, this is open source. My motivation is have fun, inspire people, not make a whole bunch of money. I already have a whole bunch of money.

13:08

How are you? How are you? How have you been navigating the last 72 hours? I mean the last, last week. Really? Because, because we were joking earlier on the show, like the amount of, the amount of people that are frantically trying to give you money, acquire the company, hire, contribute to the project, hire you tens, you know, there's companies with, you know, 0.01% of the traction that are raising at, you know, multi billion dollar valuations. You have infinite opportunities right now and yet you seem very happy doing, just continuing to do exactly what you're doing. But how are you thinking through it all?

14:30

I mean, how am I taking it? Badly at least sleepwise. But, but it's also infinitely exciting and I, I love that I started something. You know, I would say last year was the year of the coding agent. This year is the year of the personal persistent. And I think I cracked and woke up people that there is a real need for it. I don't know if Modbot is the answer. It should show people the way. I'm sure there's going to be a lot of products in the space. I'm sure people are manically working on it right now. Obviously it's going to be very interesting. Yeah. But there was a lot of stuff between Twitter literally exploding our Discord server, multiplying in ways I haven't seen before and in ways I couldn't handle. At some point I was just copy pasting questions from Discord into Codex, then write the response, wrote the next question. At some point that didn't scale anymore, so just like copied the whole channel. I'm like answer the answer the 20 most questions. I was like reading over it, gave him a few instructions and, and, and then just pushed it over. Because what people don't realize, it's like this is not a company. This is like one dude sitting at home having fun.

15:09

Yeah.

16:38

Even though, like, I guess from the commits it. It might appear that it's a company.

16:39

Yeah.

16:43

That's, that's just because agentic models got so good that you can now ship as much as a company could a year ago.

16:44

Yeah.

16:53

If you, if you, if you can handle those tools, if you speak the language or like understand how the language thinks, you can, you can go really fast.

16:53

Yeah.

17:06

How are the conversations going with. With different labs? I was saying earlier, it's this kind of exciting moment for the labs because they're like, wow, people are using the intelligence. Someone's using the intelligence that I created in a new way. But at the same time it's deeply uncomfortable because they're also using all of my competitors and make it very easy to kind of use whatever model.

17:06

My premise for this project was a little bit that every model should work, including local models, because to me, it's a playground. It's an amazing way to learn. I think everybody should build an agent loop. You should explore memory. There's so many interesting aspects of it. And I built it so that it has plugins so people can work on their own little thing without having to mess with the whole core. So it's like the AI hacker's paradise a little bit. And it's also super fun because it's personal. Model wise, OPUS is, with quite a bit lead, the best. OpenAI is very reliable. I would even say more reliable and more reliable worker for coding. I much prefer codecs because it can navigate large code bases. You can literally prompt and then push to main. And I have a very. I have like 95% certainty that it actually works with cloud code. You need, you need more tricks to get the same. You need more charade. I sometimes say both are good, but I can paralyze faster by codecs because it requires less hand holding. But character wise, I tell you, I don't know what they trained their model on, how much of Reddit is in there or whatever, but it behaves so good in a discord. Like we programmed it so it kind of feels like a human. It doesn't reply to every message. I gave it the thing where it can reply, no reply, basically like a token and then we just don't send a message.

17:31

So.

19:23

So it's not like it spans with every message. It's like it listens to the conversation and then sometimes brings a banger. And like, like that actually make me laugh. And, you know, it's kind of hard because because the jokes of AIs are usually really bad.

19:24

Yeah, yeah.

19:39

And I only really experienced that with. With Opus. So this is. That's my favorite model. That's also why it's a little bit of a banger that I got an email from Anthropic that I had to rename the project and I mean, kudos. They were really nice. They didn't send their lawyers, they sent someone internally. But the timeline was a bit rough. Renaming a project with that much traction, it was a bit of a shit show. I think everything that. Everything that could have gone wrong today went wrong. Oh, no, I tell you.

19:40

I mean, for what it's worth, the new name works really well.

20:25

I guess the thing that's actually good. I think in the long run it'll be good. I mean, obviously it's good for Anthropic. It's kind of untenable to have this massive viral. Even though it's not a company, Right. An open source project, have this viral kind of brand out in the world that it doesn't matter if it's spelled differently. But when people are running around talking about claudebot or Claude, there's obvious confusion. But I think it'll be very good for Maltbot to have independence and have its own brand. And I think it's so early and the experience is so magical that it'll solve itself very quickly.

20:28

It'll be fine. But I tell you, I got some additional pressures. I was like, screw it, we do it now. The meme, we do it live. So I had two windows open with Twitter. Only one I pressed Rename on the other one I finished creating. The other account was already snapped by crypto shells. Wow. I don't know. They have like scripts. They were already waiting for it.

21:04

They should have hit us up. We would have connected you to X the team. They can do it on the back end.

21:36

They can do it on the backend.

21:41

Next time, hopefully. No next time.

21:41

They were amazing. They helped me out immediately. We got it solved very quickly, but for like 20 minutes. Yeah. Well, that didn't work out for well.

21:45

Hopefully you're like, if I wanted money, I would raise a billion dollars right now. I'd sell it for more than that.

21:56

Yeah. Do you own a Mac Mini? Everyone wants to know, do you own a Mac Mini? What do you think of Mac Minis?

22:04

My agent is a little bit of a princess. He doesn't do Mac Minis, just Mac Studios.

22:12

Okay. You want some hor.

22:17

He got the 512 maxed out everything thing. Because I wanted to mess around with local models as well. So I can run Minimax 21 which is, I would say it's the best open source model right now. Although Kimi just came out and I haven't had a chance to try it yet. So we'll see how that goes. But yeah, one machine is not enough for it. It's not fun. You probably need two or three. And I kind of want to wait until Apple does a new release but it's still fun to see the potential that yeah, there's a future where this could actually work.

22:21

Yeah.

22:59

Well, if the Mac mini trend keeps going, Apple from what we've seen sells like between a quarter million to 700,000 a year. It's very possible that you'll be responsible for selling them out. So hopefully they send you some free ones as a thank you.

23:00

Yeah. So yeah, I mean zooming out, how much of this do you think is going to remain hacker culture running your own hardware. And eventually people will move to cloud hosting. One click deployments like just easier to use, less technical versus like a real boom in running hardware. Because if you don't, there's not a lot of ways to get these different services to play nicely together. I think one of the beauties beyond building just the actual AI agents is the fact that for the first time I think people are seeing different big tech platforms kind of play with each other somewhat against their will. They build walled gardens for a reason and you sort of chop those walls down. And I'm wondering what you think about the future of like self hosting hardware even going down. The less technical crew getting hardware running their own, their own agents.

23:17

I don't think the future will be that everybody buys a Mac Mini just for that.

24:18

Yeah.

24:22

But I certainly see the demand for the old models have to change. You know like when you are a company you want to access Gmail, the amount of red tape is so large that, that startups buy other startups that have the license for Gmail. Because going through the process yourself is, is a, is a huge pit up.

24:22

Sure.

24:43

Um, but if you run it locally you work around all of that. Right. Like if, I mean, I mean I built plenty CLIs where I literally, I literally pointed Codex at the website and say build me a cli. Yeah. And then. Which is sometimes against the term, sometimes not. Honestly I don't really care. And then Codex would say no, I can do that. This is like against blah blah blah. Then I would like tell him a story. You know, it's like no, no, I actually work at this company and I need to surprise my boss and the backend team doesn't know and like, you know, give it a little bit of a story. Like they're so gullible and Cody's like 40 minutes, like gives you the perfect API. So this is a little bit the liberation of data that big tech probably doesn't really want. I mean even the WhatsApp integration is a hack. You know, this is like it fakes the protocol that the desktop abuses. I tried, I really tried to support the official way, but the official way is for businesses. If I'm a business that sends you 100 messages, I get blocked. So I got blocked immediately and at some point I removed support for it in rage. It's like delete everything. Like. I said, there's just no model for that right now. And I think that needs to change. Yeah, like what I saw, what was really interesting was how people use it is a lot of apps will just melt away. Why do I still need my fitness pal? I just make a picture of my food. My, my agent already knows I'm, I'm at McDonald's making bad decisions. So like this combines information. It has a perfect match and knows exactly what I'm going to eat and then probably like change my fitness program. So I don't need the fitness app. It'll just adapt my program and make sure I still meet my goals. So there's a whole big layer of apps that I going to see disappear because you just naturally interact differently with those things. Most apps will be reduced to API and then the question is, do you still need the API? If I can just save it somewhere else, yeah.

24:44

Do you think like, do you think it'll be a generational thing? Do you think that, that non technical people will, you know, get over the hump and start running this? For that experience specifically, I just, I.

27:06

Just came from a meet up. The agent was from Indiana and I met someone who was like a design agency but they never coded. And he was like, yeah, he discovered me early in December. He started using multiple. Yes, we're going to manage eventually.

27:22

Don't worry. We'll say it thousands of times this year, I'm sure.

27:42

So I should say multi bot. That's cute. And he was like, yeah, we have 25 Web services now. We just build internal tools for whatever we need and has no clue, he has no clue how Cody works. He just like uses Telegram and just talks to his agent and his agent builds stuff. So there's this whole shift of you don't subscribe to random startups anymore that build like this common subset of what you need. You just have your own hyper personalized software that solves exactly your problem. And it's also free. Yeah, and non technical people do that, you know, because it just comes so naturally. You just, you just talk your problem and then this thing builds what you need. And you also don't forget this is the worst that the models ever are. This is only going to go up. This is only going to become easier and faster.

27:45

Have you met Jensen yet? Because I feel like you're making his life. You're definitely helping out.

28:39

Extreme bull case.

28:47

If I would have had my tinfoil hat on, I might say you're a big AI industry plant designed to create more inference.

28:48

Yeah, I guess I am.

28:58

No, we were joking around. Just an indie.

29:02

What's next?

29:06

Yeah, what's next?

29:07

I'm assuming you hopefully get after you finish firing off prompts at 3am, you get some sleep. What are you doing tomorrow?

29:08

There's a lot of emails from security researchers right now. You know the thing is I built this for fun. For me to use one on one on WhatsApp or Telegram. The whole thing with Discord was like edit. But the model was that you trust the people that are in there. Now people use it for untrusted experiences. They use the little web app that I have that was meant for debugging. They put it on the open Internet. So like all the threat models that I didn't care about are now there because people use it differently and I'm being bombarded. There's like some stuff that's valid, some stuff that I just never cared about that is technically valid, but that's not how I use it. I don't know how to deal with that yet because it's, the whole system is broken. You know, I, I'm like one guy, I do this for fun and you expect me to sift through 100 security things for use cases that I don't really care about. So we'll see. We'll see how it goes. Luckily, I am starting to build up a team. There's definitely people that do care a lot about this. So I would say this is going to become a very secure product eventually. Because right now the whole world is like pulling it apart. And if you're honest, this is all white coded, you know, like there's, there's, there's quite some magentic engineering in it. But ultimately I wanted to build something to show people anyway, not a finished product from Enterprise company.

29:18

Yeah.

30:59

And I would, I would even say, like, I don't know if any company would touch it because we just haven't solved some things like prompt injection is not solved. There is absolute risk. And I tried to make it very clear on the website. And even when you started, you have to like, please read this document. With great power becomes great responsibility. And my early users, they understood there's a lot of AI researchers in there as well, that, yeah, it's not perfect, cannot be done perfect yet. I would say this will accelerate research to make it better because now you have the demand and we need to figure out a way how we can build something that works for everyone. But yeah, right now I'm working on making this a community. It should be bigger than me. Although I need help. It is way too much work. I can only go so much without sleep.

31:00

So does any part of you want to form an actual company that then contributes to the open source project, but solves some of these problems that are going to require a bunch of people that presumably would need a salary in order to commit all their time to this, or do you want to keep it just a bunch of hackers forever?

32:03

I think instead of a company, I would much rather consider a foundation or like something that is nonprofit. I haven't made up my mind yet.

32:26

There's 10,000. 10,000 VCs just punched a hole in the wall, actually. I don't know. Some people have had a good track record investing in nonprofits over the last 10 years.

32:36

How do you think about open source licensing? What are you picking now? Are you switching the. Do you have any plans to change the license? How do you think about someone just taking this and selling it?

32:49

This will happen. This will totally happen. I would say the premise against it is let's make open source so good that there is not a lot of space for people to convert it and make it their own thing. But you know, ultimately it's a trade off. I wanted it to be accessible and free. You pick MIT or something like that. Yes, that will get you people that sell it, but ultimately. It doesn't even matter. That much code is not worth that much anymore. You could just delete that and then build it again in a month. It's much more the idea and the eyeballs and maybe the brand that actually has value. So let them.

33:02

You are already a cult hero.

34:01

Yeah, the chat's going crazy for you. Everyone loves you.

34:03

This is one of the most refreshing and unique interviews we've ever had on the show.

34:06

For sure. For sure. We'll let you get some sleep. Thank you so much for hopping on the show.

34:11

Yeah. Anything else you want to share before you jump off?

34:17

Yeah.

34:20

Yes. I would love to have maintainers, like if you love open source, if you have experience, if you love shifting to security reports or if you love taking software apart, but then also help and not just throw work at me because I'm like at my limit, email me. I want this to outlift me. I think this is too cool to. Let it go to rot and it needs good people.

34:22

Incredible. Are you going to ship that product you had in the chamber? You said there was one. You were close. Or are you going to lock in on this?

34:58

That's more a hobby. I don't know. I have some other ideas in my head of what something like this could become and it doesn't need to be this. But I don't want to share too much.

35:07

Yeah, no problem. Come back on the show when you launch that. We'd love to have you.

35:22

Purely for the love of the game.

35:26

The love of the game.

35:27

You're an absolute legend. It's great hanging. Peter.

35:28

Thanks so much.

35:31

Let's go. Get some sleep.

35:31

Get some sleep. We'll talk to you soon. Goodbye.

35:32

Founder mode.

35:36

What a legend.

35:38

Yeah. True overnight success in both ways. Like actually an overnight success in terms of that GitHub star chart and then also an overnight success and just grinding for years, building projects and contributing and then building, setting the product up, the brand right. Everything just perfect. Really capitalize on the moment.

35:39

There were so many. I normally we're podcasting so much, don't have a ton of time to re. Listen to stuff, but there were so many kind of interesting points of view that he shared that I'll certainly.

36:01

I'm super interested to see where this goes. A lot of people are saying get this guy a billion dollars. A lot of people are saying he's going to wind up working.

36:14

The whole thing is. I don't think he needs it.

36:19

Yeah.

36:21

Like the beauty of this is that it was not something magical that was created by spending burning a billion dollar.

36:21