Speaker 1: 0:00

The number one most difficult thing about being a founder non-attachment to outcome. Something will work because it was created by someone more capable than yourself.

Speaker 2: 0:22

Welcome to another episode of the Accelerometer. Today we've got Isu, who is building an AI tech startup. Now, Isu is a very special guest because he was one of the first people I met when I moved to San Francisco, and within a few hours of talking we decided to throw an event called AI in the Park, morphed into AI on the couch, and in my opinion, it was the best event in San Francisco. I don't think there was a better event.

Speaker 1: 0:50

There was, I remember when that was the event to go to, and that was like I was like wow, you really can just get up and just do things in the city, and that's one of the things that I just love about San Francisco.

Speaker 2: 1:03

Oh, it's really wonderful yeah.

Speaker 1: 1:05

So how's it been? It has been. It's been pretty phenomenal. Yeah, I feel like the city is like really welcomed me. I feel like I've really found my like my tribe here, and I do feel like you know, also just moved to San Francisco. I'd been here for. So when we met I'd been here for about a year give or take, but I had just really started to kind of like accelerate in Accelerate. That was not, that was not intentional.

Speaker 2: 1:32

You even know it's on everybody's minds. Right now Everybody's saying we're accelerating this or accelerating that.

Speaker 1: 1:36

Yes, I actually behind the curve because I just found out what the E slash ACC means at the end of everyone's you just learned about that? Well, because I try to actually not be on Twitter, Right? You know why? It's because I do everything, like my whole, my productivity, my mindset, everything just comes to a complete screech after I spend like 10, 20 minutes scrolling my timeline.

Speaker 2: 2:01

So where do you get your your news? Where do?

Speaker 1: 2:03

I get my news Um talking one talking to other people to when things are like I have, like, enough people in my sphere, right, which is very nice being an SF, because it's just like in the air where, like, if something is big enough, it'll find its way to me, right, like yeah.

Speaker 2: 2:20

Okay, well, that was the last big thing that found its way to you. Uh, gpt-4 vision. Okay, so you just heard about it from friends. You're just sitting around there like, by the way, gpt-4 vision.

Speaker 1: 2:30

That was actually a very funny instance because, okay, that one I was actually on Twitter and it was like that was my first time on Twitter for like four days and it was because me and my co-founder were testing something and we needed to grab a tweet to do it. And then I was scrolling my timeline I was just like wait, oh, oh, they did it, it's, it's out. Okay, all right, let's go. It was like within 30 minutes of it being announced Wow.

Speaker 2: 2:53

Yeah, that was even better than Twitter. Yeah, I feel like I've got high-latency news.

Speaker 1: 2:59

I feel like I found out about a day after it came out, oh really.

Speaker 2: 3:03

So news was like by the way, are you multimodal? Yet I was like what?

Speaker 1: 3:07

Are you multimodal yet? Oh my God, that's going to be the new RU-64 bit, yeah.

Speaker 2: 3:13

Oh, and when did you decide to build a startup?

Speaker 1: 3:16

Oh, I mean that traces back to like me being in high school watching Steve Jobs walk across the stage.

Speaker 2: 3:23

I'm like building the startup.

Speaker 1: 3:24

I was like I was the 300 billion dollar company. I, that's the startup, the billions and all that. That was not something that I understood or had a concept over the time I just was like I was just.

Speaker 2: 3:36

Steve Jobs was a bad ass.

Speaker 1: 3:38

Well, it was just like wanting to make things that help people and like and his ability to like storytelling captivate, and that was yeah.

Speaker 2: 3:44

Wow. So you were like one of the first iPhone users just sitting there, like this is amazing.

Speaker 1: 3:49

Oh no, I was an Android boy at the beginning for real, and what actually happened was so I was like I was coding. I started coding in like middle school-ish and then got an Android phone and then was building apps for that, and what actually happened is that eventually I fell in love with design right and as I was falling more and more in love with it, I realized that I could not design good apps on an Android phone. I needed to switch over. I was seeing what people were making on the iPhone and that's actually what got me to switch the iPhone.

Speaker 2: 4:23

How did you start programming? What was the first thing you ever program?

Speaker 1: 4:27

That's a funny question. It was an app for the Mac called Talk to the Mac.

Speaker 2: 4:34

Talk to the Mac. Yeah, it was. You've been ahead of the AI curve ever since.

Speaker 1: 4:40

It was. It was a chat app where you talk to your Mac. It was no networking or anything involved and basically it just had like pre-program responses to different like things that you put into it.

Speaker 2: 4:51

And you're getting money for that now.

Speaker 1: 4:53

You know, basically. Gotta revive it. If I knew the words VC, I just heard of a. I have a friend who just taught me that she runs a summer camp right For high school kids teaching them to become VC ready. I did not. I'd never heard VC ready in high school. Honestly, it's incredible.

Speaker 2: 5:13

I feel like that was my entire childhood. I went to preparatory school Like life was preparing me to be VC ready. I'm still not fucking VC ready, Like how does that even mean?

Speaker 1: 5:21

Insane because I mean, I went to public school. I did not. The idea, the words VC were not in my vocabulary for a very long time.

Speaker 2: 5:29

It's just like so explicit, like we're going to prepare you to beg for money.

Speaker 1: 5:36

I mean for real they're calculating, like you know, like customer acquisition costs, like LTVs in high school and like I'm more power to them. Honestly, like like you know, I can definitely see, like the angle where some people might be like, oh, that's kind of like you know, like like either at Lidus or like this or that, but like, honestly, like let's do more to prepare kids for like like being able to like grab life by the horns later on.

Speaker 2: 5:59

Yeah, I like that and yeah, so.

Speaker 1: 6:01

so, yeah, that was the app. The one quirky thing about the app, though, is that, um see, I was. I was bad at algebra when I was in middle school, so, um, I did not know what a variable was. So, all of the apps that I made back then and I made a couple, I made that, I made a racing game, I made like a map, I made a couple things they all had this giant window that would run in the background of whatever you were doing. It was just a matrix of text fields that were constantly changing, because I knew how to put values into and out of UI elements, but I did not know how to set a variable.

Speaker 2: 6:35

Oh, I love that. That's so, hacker, I want to play with this app.

Speaker 1: 6:41

I doubt it would run on a modern back, but literally you'd be playing this like racing game and in the background it's just like hundreds of text fields with like the speed changing and this changing. I, like I said one day, I learned what a variable was and I was like, oh my God.

Speaker 2: 6:59

Like I, I think you would feel like such a badass playing this game.

Speaker 1: 7:02

Oh, my God, you know what? Yeah, I made all the graphics and Photoshop and everything. Yeah, I just imagined like I remember like I think it was like an eighth grade or something, when it clicked Like like how to make a variable, and I was like, I was like is anyone else seeing this?

Speaker 2: 7:19

Oh yeah, yeah, so I want to try something that is a variable list programming.

Speaker 1: 7:24

Yeah.

Speaker 2: 7:24

Yeah, honestly, those were some variable list programming.

Speaker 1: 7:29

I think that nobody would show up, just absolutely nobody. Um we're going to market it.

Speaker 2: 7:36

That's true, that's true, it's all about the marketing for the variable list hackathon.

Speaker 1: 7:40

Absolutely. Yeah, this is actually it. You did not realize, but this entire thing is actually an ad for the new variable list hackathon. Yeah, the next one after that is the computer list hackathon.

Speaker 2: 7:52

I mean, I feel like so many hackathons, for the computerless hackathon was the last time you went around hackathon and you saw people sitting there with computers. They're networking events now, aren't they?

Speaker 1: 8:00

They are there, they're. They're very networking event inclined, you know, these days. But like they are, there's some good hackathons out there. There was the, there was the outside alums hackathon outside LLM outside LLM. You're not sidelines.

Speaker 2: 8:15

Yeah, it's exactly what outside lands, but harder to pronounce.

Speaker 1: 8:19

It's like outside lands harder to pronounce, um, but everybody was like hacking on music related apps Outside. That's what they have to call it next year. Oh, that's cool. Yeah, I mean, you know that you were the very first person that I ever heard the word LLM from, really, yeah, oh, that's incredible. Like I've known GB3. I've known GBT, for I'd been like using these, but no one had ever referred to them as an LLM before.

Speaker 2: 8:45

I don't know. Llm is now like the old term. We've given up on that.

Speaker 1: 8:48

Oh, what's? What's the new term?

Speaker 2: 8:49

Columbus.

Speaker 1: 8:52

Is that because like?

Speaker 2: 8:54

And you don't want to. Them are large.

Speaker 1: 8:56

I did see that, like that new model that Apple is putting into the, into the new version of iOS, it's definitely not a large LLM.

Speaker 2: 9:04

It's an SLM, it's a. It's a slim. You gotta be model inclusive. It's a slim that that is best-branded. Interesting what brought you to where you are today. So you see, you at some point fell in love with design, right? So you spent a few years designing things, right?

Speaker 1: 9:25

Oh, I designed Things, many things.

Speaker 2: 9:28

I guess the ability to serve, you have to get back into programming, hmm. So how was the journey been?

Speaker 1: 9:35

So it was. It was easy for me because I never left programming. Programming to me it was like it was like the water, even what so like? For example, for a while I was designing at Mailchimp. I was the. I led product design for marketing automation and even while I was doing that, I was building design tools so that the like, so that non technical designers on our team could do things like Embed data into their prototypes and stuff like that. Right, it was like I Would like I would.

Speaker 1: 10:05

Occasionally, when I was working on the mobile design team before that, if, um, if I was working with engineers and then there was something I designed that they were having trouble implementing, I would go and implement it and be like, okay, let's like do thing like I, like I. So it was never like a homecoming for me. It was more just like a okay, like roll my fists up, all right, this is what, like I'm going at now. Yeah, and it also means the design for me now looks a lot different. My design, like workflow, doesn't really look like me designing in figma and then moving over to um, moving over to code. I just designed straight up in code now which.

Speaker 2: 10:40

Designing code. Yeah, pure like CSS or, like you mean, like even lower level um, for me it's primarily like tell when.

Speaker 1: 10:47

But yeah, like tell when didn't react. Yeah.

Speaker 2: 10:49

Yeah, nice, I do the same thing. Sometimes you feel like it doesn't go well for me.

Speaker 1: 10:53

Oh, it's like I would say tell, tell.

Speaker 2: 10:55

When CSS was a big game changer, I kind of made it possible, but um, but yeah, yeah, no, I was designing like yours, like handwritten CSS, a race, all the properties of for the scratch.

Speaker 1: 11:09

Oh, I, I cannot say that I would do that myself, but I actually did when I first started doing this, because so, actually, when I first left my job to build startups, I was building an app that made apps right. Um, so that you could make apps without knowing how to code. That was the idea.

Speaker 2: 11:27

I know code app machine.

Speaker 1: 11:28

Yes, yeah, but the problem is that I myself Did not know web engineering. I'd only done like native engineering on mobile and desktop. So I was learning how to Code for the web, and my first project was a web app that spit out other web apps. Um, that was also pretail winner, like for me, pre all that.

Speaker 2: 11:48

So I was like I gotta ask why didn't you build this in a web app that spit out web app builders?

Speaker 1: 11:55

You know what? I think that I see what you're saying now and I think that I think we could actually go further. I think what I should have done is build a web app web app builder that spit up web app builders that would spit out web apps right, and today we just called the GPT exactly today.

Speaker 1: 12:10

We would just call that gbt4. That's actually why stopped working on it is because I use gbt3 for the first time and I looked at what I was working on. And I looked at gbt3 and I took my, the json format, the app definition format, I put it in gbt3 and then I would ask it like, hey, make the button on the home page red. And then it would just do it and I was like, oh, everything is about to change and I just, I just stopped working on yeah, I was doing Literally everything. I was like none of this is going to be the same in a year. There's yeah, yeah.

Speaker 2: 12:44

Is it that different from a year ago? Is?

Speaker 1: 12:48

Software that different from a year ago. I don't think that we've even begun to scratch the surface. I'm actually kind of surprised how Similar things have been like. We're seeing a lot of like, like, for example, there's chat gbt, that's, that's big right, and but, um, uh, but like, aside from chat gbt, which is kind of like the raw usage of these models, um, a lot of what we're seeing from. And then, of course, like the image generation stuff and the video generation stuff, that's cool too. Um, a lot of what we're seeing from software companies Is still like attaching chat to things or like an auto generated title here or like an automated suggestion there. Like, I don't. I still think that we have not really seen a ground-up re-envisioning of software categories in the Image of the llm. Yeah, I still think that the wave of ai native companies is Yet to come.

Speaker 2: 13:43

Yeah, it's a superficial what's going on at the moment.

Speaker 1: 13:47

I will like I mean, obviously I don't know what like people are working on behind the scenes, but like, so far I don't think we, I don't even, I can't even see We've spread this scratch the surface yet of what these things can really do it in the meditation, in practice.

Speaker 2: 13:59

What out there currently excited.

Speaker 1: 14:02

What other currently excites me? Um, the agent work. The agent work does excite me. I think that's incredibly cool. I think that that is actually a step towards, like a re-envisioning of how we interact with software.

Speaker 2: 14:15

Yeah, did you go to the agent hackathon?

Speaker 1: 14:18

I did not go to agent hackathon. I've been pretty Leaning away from the hackathon recently, yeah.

Speaker 2: 14:23

Yeah, can imagine it's a lot of work being a founder.

Speaker 1: 14:26

It's like I mean there's a whole lot of a whole lot of stuff. I don't also just like find myself getting into a good flow state hacking in my room, but yeah, but what were you at? The agent?

Speaker 2: 14:35

I was in a hackathon, it was just an agent meetup. It was like two nights ago here even.

Speaker 1: 14:40

Oh, okay, yeah, no, no, I did not go, we were you at the agent.

Speaker 2: 14:44

Yeah, I went there and heard the state of the art about the agents and how they're still unable to do things and I was like, did you see progress?

Speaker 1: 14:53

Did you see Anything that excited you, not just limited to?

Speaker 2: 14:58

here in general, in life in general, yeah, in being, one might say, I Think rope scaling is really cool rope scaling.

Speaker 1: 15:08

Yeah like scaling a mountain side with a rope.

Speaker 2: 15:10

I mean that's also pretty cool. But no, the rope scaling laws paper where they're where they're able to take a smaller model and Increase the size of its context without losing accuracy Like that's really exciting. I think we're gonna see a lot of improvements on model context with this pretty soon.

Speaker 1: 15:28

Okay, well, how big are we talking like these? Improvement in context.

Speaker 2: 15:32

They checked on 8k context windows. Okay, yeah.

Speaker 1: 15:36

Yeah, because, like I mean, I know that GBT4 has the larger context windows now, like Is it 32k?

Speaker 2: 15:43

and then, like Claude has like a hundred K now but when was the last time you used 32k or a hundred K?

Speaker 1: 15:49

We actually do use a lot of tokens Like because I mean we're putting full transcripts. That's it work. Yeah, actually, yeah, yeah, yeah, yeah, it is actually pretty good. They're using Claude. Yeah, we're multi-model.

Speaker 2: 16:02

Yeah yeah. This was actually the hardest thing for me about becoming a founder was learning, like, how much the aesthetics matter. To Like when raising money, when going and talking to a client, like at least it needs to look real. It doesn't matter if your tech is the best tech. It needs to look like your tech is the best tech.

Speaker 1: 16:24

Yeah, you know, it's something that like, it's something that you can actually see across like other industries, like you know, apple is the obvious example. Nintendo is actually another one that comes to mind from me where neither companies well, for a long time, apple now is like Definitely putting out some like insane stuff on the processor front and all this, but, like, for a while it's not that they had the strongest processors or the most memory like or the like Nintendo systems have the best graphics or whatever. However, they definitely had the most unique of the most unique experiences that kept people coming back right, yeah, and it turns out that's what Right play like they were playful.

Speaker 1: 17:00

They were both playful and they were both giving you something that you could not get from anybody.

Speaker 2: 17:05

It's kind of funny, though, because Sony, like they were, there were what like three game systems on the market Sony, xbox and Nintendo is a dreamcast to and just treat there was the dream.

Speaker 2: 17:16

Yeah, it's no longer recipes, but Sony and Xbox. It's not like there were that many other people for them to be just like everybody else. They were just like the Xbox or just like something, yeah. And yet that similarity Like makes it just so much that like I guess they were also just like themselves, like every every new Box that they put out was just like a faster version of the last box.

Speaker 1: 17:42

It's just more of them right.

Speaker 2: 17:44

You know? And then Tendo was like what if these boxes could fly?

Speaker 1: 17:48

literally, they were just like what if the user could fly?

Speaker 2: 17:51

I mean like what if you could take the box apart and take it with you on your? Your friends could mix and match your boxes.

Speaker 1: 17:59

I remember seeing the weird moment. I was a kid I was like I'm sorry that thing could do.

Speaker 2: 18:07

Also, you throw it at people. I was just expecting Too hard and here's a rubber casing, because if you do throw it too hard, yes, also, if you break your TV, limited liability work.

Speaker 1: 18:19

But for real, like I like, I Was just expecting like a GameCube Pro and then I was just gonna not buy it and then they came out with that thing and I was just like mom, mom, we gotta iPhone.

Speaker 2: 18:32

We, where's our iPhone? That, like I mean, we already have phones that fold and I I mean I wanted iPhone that I could like strap to the top of my head now that's the kind of thing that would excite me to see like around here people working on, like you know, new computers.

Speaker 1: 18:49

I think that that's actually like what we're really Like. Well, that's what the promise is here. Right, like with a lot of this and like, and perhaps that maybe the Innovation, the revolution, will be on the hardware front. Maybe it won't be like. I've not necessarily been a hardware centric person in my own life. I've been more software when it comes to what I build. But, uh, like there is Such insane potential for a reimagining of how we work with our content and how we put things out there that, like I I don't it just fills me with such like Intensity. I was such passion right to be able to bring that about and then also to be the bridge Right and to be surrounded by other people like yourself and everyone else around here. There are the bridges to this future and actually helping people unlock their creativity with these things.

Speaker 1: 19:43

You know what I think is the most interesting thing about like, like founding and then like operating that like.

Speaker 1: 19:48

I feel like people some people kind of talk about but you don't really like get it until you go in and try to do it is how important it is to manage your own psychology when you're doing this kind of thing right? I read this paper by like I think it was like or this blog post by Ben Horowitz a few years ago and he was like the number one most difficult thing about being a founder is managing your own psychology and, like, I always find myself coming back to that and, honestly, like every year, I just look back at the past year or even the past six months and find that like it's just truer than ever right From things like you know, like needing to, like understand one's own like quirks or how one's own mind works, and then finding workflows or sleep patterns or patterns for working with people that allow you to operate at your best right, because the thing is, at the end of the day, you're like we're leading teams here that are then, like you know, looking to us.

Speaker 2: 20:43

Everything you do is reflected like magnified. Yeah, you're like. You know, my eye is kind of itchy and suddenly the entire team is like our eyes are itchy. What's going on?

Speaker 1: 20:54

Yeah, yeah, if you, if you don't have like a roadmap together for like a week, the team does not have a roadmap together for a week, and that means the concept doesn't. That means that the idea doesn't. That means that your investors don't. That means that your community doesn't. That means your customers don't. That means what's going on.

Speaker 2: 21:13

You're a leader. You're literally, but it's not just like it's not you turn on being a leader, it's, it's. You're always a leader, and that that is. That was one of the shocking things about like being a founder is that you couldn't turn off being a leader. You're either always a leader or you're never a leader.

Speaker 1: 21:30

Yeah, what does that journey been like for you?

Speaker 2: 21:33

It's been a long and tumultuous journey. We've been through a lot of ups and a lot of downs, and right now it's going really well.

Speaker 1: 21:41

I'm glad to hear that.

Speaker 2: 21:42

Yeah, you know, it's been a process. I guess learning to manage my own psychology. That's one way of looking at it. I think another way of looking at it is learning to get in touch with, like DGF yeah, I don't give a fuck.

Speaker 1: 21:59

Yeah that's a big part of it, really.

Speaker 1: 22:05

Yeah, actually you know one of the.

Speaker 1: 22:09

So there's been a couple of these things that, like you could call them like mantras, you can call them like just like ideas that have like kind of like accelerated like one's journey with themselves, and one of the things that did that for me was this idea of non attachment to outcome.

Speaker 1: 22:27

Right, and that, to me, is like one of the biggest indicators of like not giving a fuck. Right is like can you go into a situation right and like you know you, you're fully responsive, you're not anticipated, right, you have a good time right, even in like literally like the mist of, like whatever is going on. And, most importantly, even though you are trying to steer towards the best outcome for you know you're like, you know your team, your customers like, like everyone, your community, you still have to maintain this level of non attachment to the way it goes. Because if it goes bad and here's the thing is, if you're trying crazy stuff, you will fail. You will fail, it's not, you may fail, you may not feel you will fail, and you better be able to get back up and start doing shit like immediately afterwards, right, yeah, and how can you do that if you take every little bit personally?

Speaker 2: 23:26

I feel like this is also like this, the same concept of your ABC, like always be a customer. You never, never, never, never sell anything. You're just turn every interaction into your shopping around. And that only works if you're not attached to any specific interaction, if you have an abundance mindset, if you think that there's so many opportunities in the world, like if you fuck up, you can try again, like you're shopping for the opportunity to succeed.

Speaker 1: 23:56

Yes, I agree, and it's like it's one of the things that, like that I've been talking to people about recently and I've learned is actually kind of a controversial notion is that I strongly believe that luck can be engineered or, for a less technical way to put it, luck can be manifested.

Speaker 2: 24:15

Right, Well, you find enough opportunities to be lucky. You know there's some probability every time you're going to get lucky and then one day you will.

Speaker 1: 24:22

Yeah, it's a, it's probability, and the thing is that, like it's that, like there are a lot of people who start from a place of high probability in that regard and there are a lot of people who start from a place of low probability in that regard. Right, my parents were born in absolutely terrible conditions. My dad was born in the middle of a civil war in Nigeria, and yet I'm here on this couch with you. Yet we ran AI events together that like headlined here in SF. Right, we did cool shit that I'm proud of. We're building these companies.

Speaker 1: 25:04

It was essentially because they decided that, okay, yes, our luck right now may be here, because our default if we were which I think that's what it is If you just sit there and you don't do anything, what luck comes to you? Right, it was pretty low for them. They were like, okay, we're going to force it, but we're going to use all of this inside to bring it up here. Then it's not that any single opportunities went to play out, but it's that if you keep pulling the thread on the opportunities that do come out, something will work, because that's just what happens probabilistically you just stay alive long enough while fighting for it.

Speaker 2: 25:48

I feel like this is another way of saying resiliency. You've got to be just so resilient.

Speaker 1: 25:54

Resilience and persistence and, honestly, there's a term that a friend of mine came up with. A beautiful solar name is Helena Mark. She calls it punk optimism.

Speaker 2: 26:09

Punk optimism.

Speaker 1: 26:10

Yeah.

Speaker 2: 26:12

I like that. Yeah, I feel like I don't have any of that Very pessimistic.

Speaker 1: 26:17

I feel like I wouldn't be here right now.

Speaker 2: 26:19

I was a chock full of it. I have this like everything's going to fail. But we only got one life. We got to try anyway.

Speaker 1: 26:27

Mentality I actually agree with you. I have that exact same mentality, but to me I say it with a smile on my face Everything's going to fail. Do it anyway right, and if you keep doing that, things will work.

Speaker 2: 26:40

I also found that that was not a mentality that you could direct towards a team. Your team doesn't want to hear that. Your team doesn't want to hear that the shit you're doing is going to fail. You know what?

Speaker 1: 26:57

I think that it depends on the team.

Speaker 2: 26:58

They want to hear that their work can make it succeed. It is optimism. It's this punk optimism. It is something that I've had to grapple with personally, like switching around my mindset from like because this is a founder. It's like a very internal founder mindset of like it's going to fail, but I'm going to try. Every not everybody's built that way.

Speaker 1: 27:17

This is true, this is that that part I definitely agree with.

Speaker 2: 27:20

That's like a hard way to be built to I think that there is a.

Speaker 1: 27:26

There is a I wouldn't call it a middle ground because that makes it sound more like a compromise kind of situation but there is a nuance and how things can be communicated, where you're on two ones team, where you can bring about this idea of okay, look, we try things right. And when we do try things, we put our, we put our back into it. We do it because we like, we see that this is going to work out, or sometimes maybe we do it just because we think it's going to be fun right.

Speaker 1: 28:00

And we set our definition of like what work, what working is right, but working may not be this one thing being a success, right? I mean, like Apple put out, like was it the G3 cube? And like they ended up pulling it. People don't talk about them, they're pulling it.

Speaker 2: 28:15

They're going to hurt the Apple G3 cube. I'm sure you haven't. They pulled it really quickly.

Speaker 1: 28:19

It's terrible. I mean the thing was like beautiful but it like, I think it like overheated, it cracked and like and all this stuff and like and, of course, like like a lot of the other stuff that they put out like pre, like jobs, like the Newton or the camera, people are like oh well, that wasn't under the job.

Speaker 2: 28:34

So that was a camera, the Apple Quicktick, okay, yeah.

Speaker 1: 28:39

Yeah.

Speaker 2: 28:39

No one's rid of that no.

Speaker 1: 28:44

I think they made scanners and stuff too. When jobs came, he killed. Literally he killed all of it.

Speaker 2: 28:50

I mean, I guess they make a camera now Back there, one of the most commonly.

Speaker 1: 28:56

It's come full circle and now it's called eyesight. That's funny, really. It's called eyesight, that's the yeah, because they made a webcam for the Mac called eyesight, and now they just keep using the name they don't say. I just talked about the iPhone camera is called eyesight. Really Back ones called eyesight from ones called FaceTime. There's the info. I don't know if it's useful for you at all.

Speaker 2: 29:17

But there you go, I'll go and use that. I mean bar Apple, bar trivia, when I when I produce my camera and compete with it. Oh yeah, there you go. I can't give it.

Speaker 1: 29:25

There you go, there you go. Oh, you're beginning a really fun letter from Tim Cook, right, but like, but I do think that there is a way to communicate this with teams where, like, where these things can be like a, it's like training, right, like you are, you're going out there, you're putting your all into it, but if you have like, let's say, you take your team, you divide them into two and you basically point them against each other. Right, one team has to lose right. Now, technically, you all won, right, because someone won, but you all lost because someone lost, right.

Speaker 1: 30:04

But that's almost like and this is actually funny because it would like, from what I understand, they do this at Apple actually yeah, like, I believe that's like actually how the iPhone came about.

Speaker 1: 30:14

They put two teams against each other One that was like let's take the iPod and scale it up to be a phone, and then the other team that's like let's take a map and scale it down to be a phone. And you know, I think it's clear history decided which team won. But that also meant one of those teams lost, and sometimes that means doing it in the market as well, and the thing is that, like, that is like, I think imbuing that in one's team, even though that is not how everyone is made right, that is also part of doing this right, that is also part of doing this. It's like building that kind of like movement behind it, because then, ultimately, even though you're making that impact, it's still because of this idea that the way we're doing things is, long term, going to be for the end benefit of the people we're building this for right, whoever the customer is right Either the future of how, like, llms operate, or whether it's to enable an entire new generation of entrepreneurs right.

Speaker 2: 31:21

What's something that's been challenging about being an AI founder, specifically?

Speaker 1: 31:27

So the first thing that comes to mind is that we're still very much in the early days of both prompt engineering and of the LLMs themselves, right? So even with prompt engineering, even with extensive prompt engineering, things can still be like hit or miss. So you can put like QA levels behind it and then even the QA, like prompts, can be hit or miss, right? That's the first thing that comes to mind. The second thing that comes to mind is like where the LLMs are at right, they're still extremely powerful, but I mean, we're still in the extremely early days, we're still in the first Macintosh territory of building on top of these things. Right, so much is going to be enabled and while the future is being crafted right now, but there's still like a lot of quirks that you have to deal with, right, for example, context loss over long context windows that you were alluding to earlier, things like that. And then there's also the. And then there's also and I wouldn't say this is like an annoyance or something.

Speaker 1: 32:39

I think that this is actually an important conversation that we are at the precipice of right now, which is a lot of the unanswered questions around using these models to create these businesses. Right, everything from the questions around copyright, to the questions around like the like, to questions around like ethics of it, to questions around the like, the idea that it could put particular people like out of like business right, like it can take jobs, and things like that. All of these are things that I think like the entire industry is grappling with and that it's also like each of us are also kind of like grappling with as like founders, as we like build these companies. Yeah.

Speaker 2: 33:26

What's some tech? You would like to see more founders focusing on.

Speaker 1: 33:30

New computer.

Speaker 2: 33:32

New computer?

Speaker 1: 33:33

Yeah, the LLM computer Like like a re-envisioning of how we interact with software. I think that I think that we have become very comfortable with our current paradigms and it is becoming harder and harder for us to see how we interact with software from a lens, beyond a user and screen, and I think that we have now just been given the tools to blow far beyond that Right, just far beyond that Right.

Speaker 2: 34:10

So some of what, like they're not concerned at all about, like the accuracy of these systems.

Speaker 1: 34:15

Well, the accuracy of these systems is okay. So I feel like there's two fronts to talk about the accuracy of these systems. One is the factual accuracy of these systems.

Speaker 2: 34:24

Right.

Speaker 1: 34:24

And then the second is the sort of like comprehension, or like almost kind of like ideation, or neural processing accuracy of these systems. On the factual accuracy of these systems, we're relying on the content that it already knows in its head. And, yeah, I can see it making a mistake. I make a mistake. If you ask me like things like right off the top of my head, I will Some of them. I'm just gonna have to tell you I'll have to get back to you on that Because I'm not in front of the internet, right.

Speaker 1: 34:51

I don't have like my fact book in front of me and I only have like so much memory as a human being. And the way to do that is in the same way that we supplement ourselves with calculators. We supplement these LLMs with knowledge repositories, whether it's like a company knowledge repository or it's the internet itself, right, so we can use that to pull facts. So then, in that case, I tend to see LLMs less as being a source for factual knowledge itself, right, but more as like a neural processing unit to be able to take factual knowledge and then basically perform like word or syntactic engineering on it after the fact.

Speaker 1: 35:32

Then comes to that second part comprehension, like accuracy Right, and I'm looking to do that, and a big part of what we've seen is that it's not perfect. It's definitely not perfect, but it is very, very good and it is getting much, much better very quickly In terms of going where the hockey puck is going, right, like looking I'm looking towards where the hockey puck is going and not where it's coming from. That, like every cell in my body just says like that's where it's going. We're only going to see the accuracy improve.

Speaker 2: 36:07

What do you want to tell the world? What is important for you, for your life?

Speaker 1: 36:12

One of the most pivotal and most important things that I'd ever heard in my life was it was a quick Steve Jobs made.

Speaker 1: 36:23

Actually, I remember hearing when I was younger, and every year it's just like kind of meant something more and more to me. It's the idea that every single thing around you was created by something, was created by someone no smarter and no more capable than yourself, and I think that it's very easy for us to hear that and put a lot of caveats on it On two fronts. I'm not realizing just how deep that goes when I say everything from your shirt, to these mics, to the cameras, to the disco ball, to Disney, all of it. And then the other side is the caveat of being like oh well, there are people smarter than me, I'm smarter than me, you know MLK was better at rallying people than me, all this stuff, but I only continue to find that to be more and more true. Yeah, everyone is capable of making so much incredible cool stuff, and everything that was everything that you see around you was made by someone no smarter and no more capable than yourself.

Speaker 2: 37:55

That's really beautiful. I wish I could take one before I click it off. I think that. I think that is just a wonderful sentiment and you know, I'd never heard that before. So thank you for telling me that and like looking around right now like you know what I could make that disco ball but like yeah, just take a mirror and shattered it to a tiny piece of paper and stakely put it on. And like almost gives you permission to be more creative as a person.

Speaker 1: 38:24

It does. It's so funny. It almost like like there's like a that can be a permission needed. Some people are just like, born without permission. That's fantastic, right? I definitely had to learn to give myself that permission. I was not born with this idea that I could just go and do anything, right, or?

Speaker 2: 38:41

I was how you to do things and I feel like I'm fighting against myself to do things very frequently.

Speaker 1: 38:48

Exactly Like. There's like a like an inter apprehension, right, but that apprehension can actually be compounded away the more that you do the thing anyway, the more that you trust yourself to do the thing anyway, yeah.

Speaker 2: 39:05

You ever find yourself like rebelling, rebelling against authority to create. To create, yeah.

Speaker 1: 39:12

Every damn day.

Speaker 2: 39:13

Yeah.

Speaker 1: 39:14

Yeah. Do you ever find yourself rebelling against authority to create Mr Head of Anarchy? Oh yeah.

Speaker 2: 39:25

I mean, like I feel sometimes it's just so frustrating. It feels like literally everybody is trying to stop you from doing what you want to be doing. And this is what I mean when I say like being like my journey in being a founder was a journey to de-gaff. Like when I named the company Anarchy, everybody was like that's a terrible name, please don't do that. And like, because me trying to please everybody else and instead of just doing what like I fucking wanted to do, I want to call the company Anarchy. It's my fucking company. Like Anarchy is a cool name, I'm going to call it that. And like, yeah, sometimes it feels like the entire world's against you, like, even if they're not actually, it might just not like this little thing that you're doing, they don't really fucking care, because they're not you, they're not living your life, and yeah, but sometimes they literally are, they literally are trying to stop you.

Speaker 1: 40:15

You know, sometimes, sometimes, yeah, but I do also find that like there's a difference between people like trying to stop you by saying something that is discouraging or by having like their own apprehension right, versus trying to stop you by like putting physical work into trying to stop you. And for me it's like if somebody is just saying the thing, then they're just saying what came to their mind, and maybe it comes from a good place and, honestly, maybe it doesn't, but those are words.

Speaker 2: 40:48

I had an employee literally go and be like no, you can't call a company anarchy and sit one around telling people like the other companies are called out at orky.

Speaker 1: 40:56

That is more than words. That's different. They're going, that's an employee telling people that the company is not called something, that it's called. That's a different thing. But yeah, then then there's when you know people actively do things like physical things, to impede your progress. In which case, well, in which case, like you know, I feel like it's time to be the parent, as a founder, you have to go beyond the decaf to like.

Speaker 2: 41:22

You got to get the decaf but you got to go beyond it into the like no, I give a fuck, I'm doing the thing like anyway, because it happens enough. Like people trying to physically impede you, that like if you were it like you. First you have to get over your own homes, over the people who just don't like the idea, and then you have to get over the people who like aggressively, like no, I don't want you to build back, because enough people are out there. Your competitors if you have competitors will try to stop you eventually.

Speaker 1: 41:49

Oh sure, but I mean like those are competitors.

Speaker 2: 41:55

So is everybody is trying to stop you. No, competitors trying to stop you is one thing, other people trying to stop you. Local governments parents.

Speaker 1: 42:03

Okay, well, that's a different, that's also a different situation Now you're talking more like the police, the decels.

Speaker 2: 42:12

Have you ever encountered the decels?

Speaker 1: 42:14

I can. I've not heard that term in this regard, but I can assume. I think I can fill in the opposite of the X. Yeah, I came first. Yes, I do.

Speaker 2: 42:25

That is what I figured it would be In fact, that was the one of our and the couch meetups and the park meetups the first one and somebody came who is a decel and I had this long conversation with him and just like no, I think that all AI progress is bad. Why are you here organizing these AI people? It came to try to stop us. That's what he said.

Speaker 1: 42:45

Yeah, that's interesting. I don't think the cat is out of the bag. The cat's out of the bag. If we're being fully honest and realistic with ourselves, the cat is out of the bag. I do think that it is now a matter of there's regulation, like there is regulation, yeah, and you can regulate here in the US. There is the rest of the world that also has access to this technology that will not adhere to those regulations, which is, you know, also something that I don't hear discussed a whole lot in that.

Speaker 2: 43:21

So you're going to fight the regulation?

Speaker 1: 43:23

It depends on what it is. I don't know.

Speaker 2: 43:26

I hope you want to roll over, though they're going to come and say, like no, your business, we're going to regulate it away.

Speaker 1: 43:33

We're not a roll over kind of group of people. We are basically like we're here, like I haven't really talked a whole lot about, like what our company, trump, does, but at the end of the day, our whole thing is basically fostering a new generation of entrepreneurs by making their work simpler and taking away the on it, like basically the bullshit work, the work they shouldn't have to do. Yeah, I was part of it as part of something starting something from scratch, and in that regard, I mean we're going to, we're going to fight to make sure that we're able to do that for these like customers, right.

Speaker 2: 44:07

That's yeah, I think that's wonderful. That's what I would love to hear if I was a customer. No, well, I think that is a great note to end on it. All right, thank you so much for coming on the accelerometer. Oh yeah, it was really wonderful having you and seeing you again. Thanks for having me, matthew.