The Exchanges

Q What if the technology just isn't good enough yet, and it needs to be improved?

A It needs to be improved, definitely, but that wasn't even the response that I got. The response that I got was just, like, shouting and throwing things, and so I was like, something feels wrong here, but I wasn't really in a position then to pursue it, but this was always a thing that was kind of, I saw that there was a really important unification here. It also, this, the same fundamental type of technology has really transformed organ transplantation. So there, they call it NMP, normothermic machine perfusion, rather than ECMO, but it's the same idea. Um, so, 20 years ago, if you needed a, like a, Kidney transplant or liver. If the car crash happened at three in the morning, the surgery would happen at four or five in the morning. But now it gets scheduled for like the afternoon or the next day, and over 75% of liver transplants in the U.S. use this type of perfusion technology now. But like the, the systems that exist for this are like 500,000 dollars. They can only be moved by private jet. Like one of the big companies in the space, it turns out that they're like private jet logistics business is bigger than their medical device business. And it just like, there was just like clearly an engineering that could refine this. And so we looked at this and we thought like, well, what if you could refine this to the point where you could check a kidney as luggage on a United fligh…

AI assessment note: “It needs to be improved, definitely, but that wasn't even the response that I got.”

Q Maybe we start earlier because one of the things we'd love to kind of, you know, get down as a part of lore is like, how did you get so AI pilled and like all the way to the edge?

A Well, I'll tell you my encounter with LLMs, which was so, so I remember in the pandemic, um, there was someone, someone gave me an API access to GPT-III and, uh, and I was playing with it and I was like, okay, this is, this is really cool. This is, there's, there's something here that could be, could be special. But it was the kind of thing that was like, yeah, it feels like a research project, the kind of thing that Google used to release, and you're like, you play with it for 10 minutes and you stop. Uh, Chachipiti came out, and I think everybody was sort of interested in it. Where I think it got interesting was, uh, when you started to see Reasoning models and of course tools, but, but I think everything else was sort of a blip until December. Um, and the way I describe it to, to my team is like, you know, electricity was invented in December. Uh, and I think electricity was Opus 4.5 and, and sure Opus models and, and, you know, open AI models got, got better and better since then. But, To me, that was the, the, the, the tip of the spear where you could say, yes, like coding harnesses actually work and, you know, cloud code existed for probably a year before, uh, but it wasn't that, that valuable yet. And I remember, you know, during the holiday break, I was playing with it and, and it was, was pretty shocking, probably similar reaction that, that everybody here had. And I t…

AI assessment note: “Well, I'll tell you my encounter with LLMs, which was so, so I remember”

Q Why digital marketing? Like you could have done like factories for anything. Why is this the thing you decided to build the factory for?

A I think that the web is still one of the most transformational technologies. Um, sure, it's done some funny things with how we consume media, how we respect sort of media organizations. However, ultimately, this sort of democratization, the dissemination of information for the internet, to me, is a really exciting tenant. It's a really exciting space to be, even after working in this space for, for dozens of years. My approach is that Where we're at in AI, these businesses, they still need to be found. And I actually think there's going to be way more small business in the future. You're not going to have like massive companies anymore that are dominating. I think where, where society is moving is I actually think like entrepreneurship might become way more important than it has been. Entrepreneurship, especially if you're a small business, you need to be found. You need to be able to tell your story. You need to be able to represent your brand well. And I think that's just like a really exciting space for me.

AI assessment note: “especially if you're a small business, you need to be found”

Q Then fast forward to the founding of Ennisphere. It's an interesting name because Cursor is not what it is. When you guys started, you had just, um, graduated MIT, right? That was back in twenty-twenty-two? What were the first ideas that all four of you started working on back in twenty-twenty-two?

A Yeah, so the, the genesis of Cursor was in twenty-twenty-one. Uh, my co-founders and I, we had been interested in AI for a long time. Each of us kind of had our own little robot dog moment where one of my co-founders, he worked on, uh, trying to build a competitor, Google actually, uh, using LMs in, in twenty-twenty-one and, and training his own, um, and training his own contrastive models. Uh, one of my co-founders, uh, worked on computer vision in academia and, you know, some of us also worked on recommendation systems at, at companies like Google. But, uh, we were really interested in AI. In 2021, we were trying to figure out what we'd do with that interest. Do we go and work on AI in academia? Or, you know, do we go join, you know, a big existing AI effort? Or do we start our own thing? And there were two moments that really got us excited. One was seeing the first AI product start to come out. Uh, you know, GitHub Copilot was really the canonical example for us. The other was seeing work about how it looked like AI was going to predictably get better in the future as you scaled up these models. At the very beginning of twenty-twenty-two, uh, me and my co-founders, we went on a, like, a month-long hackathon, basically, and we started hacking on ideas related to kind of picking an area of knowledge work and building what it looks like as AI gets more and more mature.

AI assessment note: “we started hacking on ideas related to kind of picking an area of knowledge work”

Q Yeah. It sounds like it would, um, it would take a similar shape to, uh, creating OpenAI in that, like, do you need to be ready for, uh, people to attack you or dismiss you?

A Definitely. I mean, if you do anything that matters in the world, uh, you will have a lot of people call you an idiot or just dismiss you. The more you do, the more, the better you do, the more they'll attack you. Um, the more you kind of, like, threaten the existing state of the world, it will just continue to escalate. Um, One thing that I've noticed about many of the best ideas is the vision is clear. Like we wanted to build AGI, but the, the first steps were super unclear. Like we didn't know that we were going to be a product company. We started this nonprofit research lab for many reasons, but one of which was it, it really didn't occur to us that we were going to make a product that people would be able to pay for. Like this was, you know, it was years till we came up with the idea of ChatGPT, the API. And so, I think if the, the kind of, like, highest level vision is clear, but the first few steps are very unclear, that's not a bad thing. That often happens with, like, very ambitious ideas, and I wouldn't let that cause you to lean out. Now, you do still have to take some steps forward, imperfect, though they may be, and ours were certainly very imperfect. So there's, like, another failure case where you have this, like, brilliant big idea, and you can kind of never make any forward progress. At some point, you've got to just, like, Got some new data points.

AI assessment note: “Definitely. I mean, if you do anything that matters in the world”

Q Basically there's, um, Delaware bylaws. Is that right? Basically, you know, if you're a Delaware C Corp, you have to relentlessly pursue profit. Otherwise there's grounds to remove you. That specific principle is exactly how, uh, you know, the end of that founder's reign of that particular company happened.

A I still remember the professor being like, Cause I was like, you're not listening to me, right? You're not getting it. And he's like, wait, are you saying that's going to be me someday? I'm like, you're on a one way ticket to this exact outcome because you've adopted the so-called best practices of corporate governance of how companies are supposed to be built and run and structured. One of them of course is what's called shareholder primacy, right? This idea that if you are Delaware C Corp, the thing you make is not a beautiful living thing that creates products and, you know, delights customers and, and it's like a good, no, It's just a financial instrument for investment returns. That's what, that's all it is. That's actually a very new idea. And I think one of the things that's a big misconception for founders is they assume that this is some kind of natural law or like a pillar of capitalism going back to Adam Smith or whatever. No, Adam Smith would have been like, what the F you guys talking about? This idea dates to the 19 eighties. The professor was saying to me, he was just like, wait, so is it possible to build an incorruptible company? That's kind of how the book got its title. And I was like, well, it's a good news, bad news kind of thing. Everyone says this is impossible, that like this kind of corruption of the mission is natural. It's just, as you get bigger, as …

AI assessment note: “One of them of course is what's called shareholder primacy, right?”

Q For those watching who have never heard of a brain-computer interface, what is it, and what have people been able to do, what are they able to do now?

A So the brain is this powerful computer, but it's encased in the skull, like it is not magically connected to things, and so, um, it has these, these handful of connections to the world, and these give you the senses that you know, and the motor control that you know, but You can kind of ask like, is that, so either do we want to replace these with something else? So for example, like the simulated reality or the matrix use case. Another is restoring lost functionality. So this is, I mean, this is how they're deployed today. So if someone has gone blind, you can restore the ability to see. If they've gone deaf, you can restore the ability to hear. If they're paralyzed, you can restore the ability to move. And then you can think about structural neural engineering. And this is the, this is the thing that people haven't really, we haven't gotten to as a field as much, but looking at how, how does the brain process information? Can you add new brain areas? Are there ways to understand how the brain is like, what, What is going on either to use this to build smarter machines or to think about how to treat things like depression or addiction.

AI assessment note: “Another is restoring lost functionality. So this is, I mean, this is how they're deployed today.”

Q it's because the product just works. I don't think I have a better answer. I've always been curious, like, Why is it so difficult to make the product work? What's actually so hard about when I type in, I want a software engineer with 10 years of experience, um, um, why is it so hard to surface, like, the relevant candidate, and how have you been able to do that?

A A little bit of context on Juicebox, we do search, um, contact management, outreach, and then sync all of that data with your ATS, and of those, search is really the hardest part, and so being able to find the right person is a hard problem for two reasons. One, There's a long tail of different search queries. And so if people are using the platform correctly, no one will have exactly the same search query because there's always something unique about the rule or unique about the opportunity. That also means that there's a long tail of, um, potential filters we might need or potential ways to, to think through how we filter a segment of search. And then that relates to the depth of a profile. So if someone has previous experiences, those are really important in a search result. And it's also quite different than how search on like a sales tool might work Like, where it's all about the current company or the current role. Um, with recruiting, it's really about their full depth of experience. And so, I guess, long-winded way of saying that search is really important and is usually what drives those outcomes.

AI assessment note: “being able to find the right person is a hard problem for two reasons”

Q I think Gary's got, like, a great point on this. Um, it's basically, like, become, like, a forward deployed engineer, right?

A Yeah, just, I mean, go undercover, I guess. Like, go, go and figure out what people actually need. And, um, yeah, there are just too many examples of billion dollar startups that we got to see. I mean, I always think about Flexport. You know, here's this guy who literally became one of the top importers of medical hot tubs. Like, I don't think anyone wakes up You know, and graduates and decides like, hey, I really need to become one of the foremost, you know, import exporters of, ah, of medical hot tubs. But, you know, he did it. He, they, they did, they also, um, I think were one of the first, the biggest e-bike importer. But then, you know, basically being in weird parts in the economy, um, caused them to understand just things that, that, ah, the, the other person, you know, the sort of 1010 thousand other people who want to start startups, like, they didn't have that knowledge. And so, Sort of your ability, your, you know, if you're here, like your inherent ability already is like one part of the Venn diagram, and then the other part is just something weird. It's literally just like, where does your interest come from? Like I'm really taken by to what degree both OpenAI and SpaceX, for instance, where, uh, you know, the Genesis came from like interest and a hunch and just Like, not really any commercial intent, and yet, you know, coming out the other side, uh, that was enou…

AI assessment note: “Yeah, just, I mean, go undercover, I guess. Like, go, go and figure out”

Q Um, and, um, Is it true that you're planning a robot army of a sort?

A Whether we do it or, or, or, you know, whether Tesla does it, you know, Tesla works closely with XAI. Um, like, if you've seen how many humanoid robot startups are there, like, it's like, I think Jensen Bong was on stage with a lot, with a massive number of robots, you know, robots from different companies. I think there was like a dozen different humanoid robots. So, I mean, I guess, you know, part of what I've been fighting, and maybe what has slowed me down somewhat, is that I'm a, I'm a little, I don't want, I don't want to make Terminator real, you know. So, I've been sort of, I guess, at least until recent years, dragging my feet on, on AI and, and humanoid robotics. And then I sort of come to the realization, it's happening, whether I do it or not. So, You got really two choices. Particip, you can either be a spectator or a participant. It's like, well, I guess I'd rather be a participant than a spectator. Um, so now it's, you know, pedal to the metal on humanoid robots and, um, digital super intelligence.

AI assessment note: “so now it's, you know, pedal to the metal on humanoid robots”

Q What are you seeing about, like, so the concern around data privacy is another big reason. Like, are you seeing that as being enough, like, are people worried about giving these data sets to OpenAI?

A It's really interesting. I mean, whenever you have something so new like this, it's actually, um, sort of resets the clock on the competitive landscape again. So, you know, you almost can expect all the same things will happen again. Um, you know, just as 10, 15 years ago, cloud was brand new, and then you had cloud cybersecurity and cloud strike and all these companies sort of come out. Um, you know, we're seeing the first wave of cybersecurity companies, you know, like prompt armor. So they sort of wrap your API calls And, uh, what they actually have figured out is that for a lot of large language models, if you do any sort of fine tuning or training with private data, you can actually just speak to the model and get it to spit out your private data again, and they have a solution that stops it. So it's so interesting because, you know, it's entirely possible, you know, they're basically creating a new industry again, um, of cybersecurity for LLMs, sort of in the same way that Cloud opened up that space and created cybersecurity for the cloud.

AI assessment note: “we're seeing the first wave of cybersecurity companies, you know, like prompt armor.”

Q Uh, here we go. Oscar says, we're blue, a platform where housewives resell products from suppliers using WhatsApp. It's like Michaud for LATAM. I get that. Should we give incentives to our housewives to trust us like Uber did at the beginning? Incentives.

A So, um, interestingly enough, This is a common thing that happens where a startup thinks they know what happened at the beginning of a company, and sometimes they don't. So in the extreme beginning of Uber, what was so interesting was that it launched in San Francisco, and it launched as Uber Black, which is a really expensive product, actually. It wasn't well subsidized. And the real value is that it was impossible to get a taxi in San Francisco. Absolutely impossible. So they didn't have to incentivize people to use the first version very much because, like, the alternative was, like, you couldn't move around the city at all. Now, as they expanded to new markets, and they figured out how marketing worked, and they figured out that, like, they could accelerate penetrating into a new city if they gave away rewards, that was all stuff they did later.

AI assessment note: “So they didn't have to incentivize people to use the first version very much”

Q Let's see. So, one question. I mean, when you look back on the decade, um, what do you think they'll say was obvious in hindsight about AI that people are just missing in real time right now?

A You know, so much of the debate that happens these days is around, oh, how good are the models actually getting? And can the models actually bridge this issue? And, you know, when are we going to get super intelligence? Is that in like two years or five years? And, you know, are we going to hit a wall? And, you know, so much of that debate is like, I think, um, in some ways, uh, a little bit of a waste of time, because, you know, I think it's inevitable that we're gonna have very powerful models, and, um, you know, rather than, I think we'll look back and say, oh, all this arguing around, like, when exactly it was gonna happen was sort of, um, was short-sighted, because the reality is, we are just, as a entire human civilization on this incredible exponential, Like, you cannot look at the progress of AI over the past decade and not just be totally awestruck by how far it's come. Like, a decade ago, the best AI models could recognize cats in YouTube videos, and now, you know, we're talking to, um, you know, a digital god that can, you know, uh, I mean, we've all seen some of the hacks and some of the, some of the things these systems are capable of, and You just can't help but be awestruck. And, and I think this trend will just continue. Like these, these models are going to become more and more powerful. And so I think a decade looking back, it'll, it'll be obvious that intelli…

AI assessment note: “looking back, it'll, it'll be obvious that intelligence became abundant”

Q So let's talk about super intelligence, because that's clearly, that's even in the name of your lab. Um, what does super intelligence mean operationally inside Meta right now?

A Yeah, I think that, you know, we, a year ago, Mark wrote this, um, uh, memo about personal super intelligence, which I think actually is very similar to your concept of personal AGI, but, you know, we believe that everybody in the world, you know, all the billions of people in the world are going to have a super intelligence that is adapted and tailored to them, that is, enables them to accomplish their goals, knows their context, and ultimately is an expander of their own agency. Like, I think the thing that we think a lot about is, is agency expansion. How do we help people accomplish things that they couldn't have ever dreamed of before? And what would everyone in the world do if everything was just easy? Um, and we think about this in a, in an ecosystem way, um, as well. I think, uh, you know, Patrick mentioned it, but, you know, we don't believe in this totalizing, you know, totalitarian view of, you know, AIs that control the world. We believe that these are going to enhance this very broad ecosystem. So, you know, we believe in billions of people all around the world all having their own personal super intelligence, and we also believe in, you know, an explosion of entrepreneurship. There's two hundred million businesses that, uh, are on Metis platforms today. We think that number should go to billions with this explosion of, of creativity and using AI tools, and ultimat…

AI assessment note: “Mark wrote this, um, uh, memo about personal super intelligence”

Q non-venture funded things, it's not like a choice if you're bootstrapping about whether or not you charge for your product, right? Um, and so just like, where would you advise yourself? You know, say you have a time machine, you could advise your younger self. What would you have told yourself around monetization? It would have been like, you know what? Wouldn't change anything. Like, yeah, that's a valid option.

A It's, yeah, it's one of those things, because I, there's a lot of mistakes I made, of course, and, and so I, when I think about the mistakes, I think, well, I should have done it differently, and then there's another part of me, though, that thinks, well, with the information I had at the time, I probably still would have made the same decision, so the reality is I don't think things would have changed, and it's also hard to know if I did that, would that have actually been worse than what happened, and totally fair, so the things that I, I, I think we made a mistake on that I, I take responsibility for are a few different things. One of them, one of them is we should have dedicated a little bit of time and focus on revenue. And in the beginning, before we were funded, we were, you know, I was paying the Heroku bill out of my own credit card and it wasn't expensive, like a couple thousand dollars a month, maybe at one point. And I was like, well, I don't want to pay this. So let's like do job ads.

AI assessment note: “we should have dedicated a little bit of time and focus on revenue”

Q Who were some, some of the first hire? I mean, I see more engineers, but you know.

A So we agonized over the first hires, and I think that if you want to go fast on the order of years, actually going slow on the order of, you know, six months is super helpful because if you really nailed the first 10 people to come into the company, they will both accelerate you in the future because when, you know, the nth person comes in that's, you know, is thinking about working with you, comes and hangs out with the team, they'll just be shocked by the talent density and then really excited to work there. And then the other reason they can help you go faster in the future is if someone comes in and they're not a great fit, these people act as an immune system against that, right? And they will be kind of keepers of holding the bar really high. And so we hired very, very, very slowly at the start. We were able to do that also partially because we had such a big founding team and all the co-founders were technical, but yeah, the people we got, uh, uh, are fantastic and are really core to the company today. And Folks who bled across disciplines where we are this company that needs to be something in between a foundation model lab and a normal software company. And the models and product have to work together under one roof. And so we had fantastic people who were, uh, product minded, commercially minded, but had actually trained models at scale.

AI assessment note: “Folks who bled across disciplines where we are this company that needs to be”

Q finally have the tools, and now it's about entrepreneurs coming in and saying, how do we use these tools to make 10 X impact for businesses or to build new consumer experiences that were previously completely impossible because we couldn't program things on a Global stage. We couldn't move money at the speed of light, but are now possible. And how do we actually use that to create this value?

A I want to talk about stablecoins more. So like you mentioned, though, is that so far it seems to be outside of trading, like the killer use case, um, specifically within the Y Combinator community in our portfolio of companies. Obviously looking at this, some of our fastest growing companies at the moment, the AI companies definitely get lots of attention, but we have companies like Dollar App in Latin, Aspora in India. These are essentially neobank Services that are really built around stable coins and their growth rates are incredible. I'm sure you guys have seen this at Coinbase too. Let's start with this. Why have stable coins taken off in this way? And actually I'm curious as someone who's been around crypto for so long, has it surprised you? Did you think it was going to be stable coins all along?

AI assessment note: “I want to talk about stablecoins more.”

Q harness is, uh, you know, underwrapped still, but, like, can you tease us with, You know, I mean, I still use Open Claw. I still use Hermes Agent. You know, it's, ah, you know, these things are, I call them Ferraris that break down on the side of the road all the time. Like, is this a Ferrari that won't break down? Like, you know, tease us a little bit.

A Yeah, hopefully, hopefully it doesn't, it doesn't break down. I mean, I think we're really focused on speed. I think speed is, um, you know, for anyone that uses these tools, speed is probably the, you know, one of the most critical things. I think also reliability, like you mentioned, we want it to be extremely reliable. Um, We want to be very extensible and to scale to as complex and interesting of a multi-agent setup that you, that you want to have. Like, I think there's so much innovation that will occur even above the harness, frankly, um, in terms of, like, how to orchestrate and set up loops and, and develop, like, you know, very complex ecosystems of these agents working together. Um, uh, we want to be really extensible and, and, um, ultimately we want to just Empower people to harness this technology, because harness, ah, no pun, actually, pun not intended, but, um, but there's, like, I truly believe these, these models are already just incredibly powerful. Like, they should, they should be so powerful to fuel, you know, um, many, many points of expansion of GDP growth, and I think it's, like, up to smart people with vision and ambition to make all that happen.

AI assessment note: “we're really focused on speed... reliability, like you mentioned, we want it to be extremely reliable.”

Q Yeah, yeah, I mean, like, I, I feel in the metaphor, I'm just, I'm thinking of, like, the, the part where we eat the financial world. I'm like, what's the, what's the thing that's hanging through?

A You're seeing, you're seeing, um, Um, you know, super, you, you can go today and start writing a, something that behaves like equity or something that is a derivative or, you know, all of these kinds of financial instruments that would take you a ton of time to kind of think about and reason about and, like, inject into, into the jurisdiction, you know, any kind of legal jurisdiction in the, in the world, and you're now able to do that in a, in a, Totally different way with a whole bunch of assets that represent real value. Um, and so, like, I, I think that there's a ton of these that have very direct use cases and applications, but they're not, they're not consumer. And, and so that's, that's why you're seeing a wave of things that seem weird to Silicon Valley. They seem like, oh, this would never work. And yet, there's a ton of companies out there in the world that actually need these kinds of things, that actually think through it, like, oh, wow, like, that means I don't have to spend Hundreds of thousands of dollars to millions of dollars in legal just to understand, reason about, and conduct these transactions, and then have to worry about litigation down the road of, like, in the millions of dollars to try and make sure the transaction is safe, um, you can then turn that into, into, like, single dollars, right, of, like, running transaction fees. And, like, that is a mass…

AI assessment note: “you can go today and start writing a, something that behaves like equity”

Q Now my last question is, given everything that we talked about, if there's someone here that's studying CS, and you, you learned to program before this era of, uh, AI in genetic coding, what should students still learn the hard way, like the old way?

A So for me, I learned computer science practically. I learned it by teaching myself to code in order to solve problems. Whenever I was doing this, I was doing it to solve a particular problem that I had. So I actually first learned to code on a TI-Eighty-three calculators. Um, this was back in middle school, and, um, I ended up actually writing a guide on the internet for programming TI-Eighty-three calculators. It's still up on the internet somewhere. Um, and it was, uh, it was basic. That, that was my first language. And I, I learned how to program on the calculator, so I could just, like, get better at my math tests by, uh, by cheating on the test. So it was about something practical, you know, like to me as a middle schooler, that was kind of like the most practical thing I could think of. And I ended up getting good grades, and then I got this little serial cable to give the, you know, the programs to my classmates, and they got really good grades. And then the math got a little bit harder. Um, it wasn't something that I could solve in basic anymore, so I kind of went from this, like, you know, like, maybe algebra solver that was written in basic, and I had to solve harder problems. And, um, you know, like, once we got into calculus, I had to run assemblies so that I could write A better solver so I could cheat better on the test now that it was calculus. And so, for me, pr…

AI assessment note: “learn not just the computer science... but learn how to apply it”

Q have an existing customer base. They don't have revenue. They can like, their customers are going to be more forgiving if the thing doesn't work yet. And these seem like advantages for like new startups attacking incumbent companies with like an AI product. Um, but from your perspective, what are, what are like, what are the advantages that like incumbent companies have over them? How do you think about it?

A I think the real business has to be downstream of AI visibility. Uh, it's a valuable, there's value in it, like, you know, just like there's value in, in the SEO world, but it's so easy to do. You know, we, we did it, you know, as, as I mentioned in, in, in, uh, probably more like a few months and a few weeks, but we, we did it very quickly, um, and gave it away for free, and it's been this incredible lead generation for us. Um, and so the commoditization is going to happen real, real fast. I mean, you talk about like, you know, people talk about, Uh, hey, uh, SaaS going away because of AI. I think that's a great example of it. Um, and so, in contrast, if you look at, I think, businesses that have done very well here, so like, there is a business, I think, that is very viable here, which is what Aerops is doing, where they're, yes, they have some visibility aspects, uh, but they have, uh, a whole content generation business to help you create blog posts and other material, um, on that, you know, and, and that's the, that's their real business, and so I think all of these visibility businesses, like, You can, you can get it for free from us. You can get it, you know, you're going to be able to get it free from a lot of other places, and you're going to have to construct a real business kind of downstream. And maybe to give the flip side of it, like, um, you know, I, I think inno…

AI assessment note: “because we have an existing revenue base of hundreds of millions, we can give away this”

Q that seems like it's a big catalyst for this particular moment in time for crypto, there's lots of news around, like, the Genius Act is going through, and you've got various bills trying to sort of regulate The crypto market. In what ways have you seen regulation hold people back from being able to like really innovate with crypto over the past few years? And how is that changing now?

A Yeah, well, I think this is something that really impacted startups of the type that are coming into YC and trying to be successful, you know, and I've been working with builders over the last, you know, five, six years to try and figure out how can I help them be successful building on chain? And the thing that has been really consistent is that So many early stage builders and late stage builders ended up spending like equivalent or more money on lawyers than they were on engineers. And you know, if you're coming into YC and you're interviewing and you're like, Hey, we're going to be spending more money on lawyers than engineers. Like my gut is that the YC partners are going to say, okay, like maybe you're not building in the right business here. Like really, are you as a three person team going to be able to be successful?

AI assessment note: “builders ended up spending like equivalent or more money on lawyers than they were on engineers”

Q Probably throughout that period, a lot of people were saying, you know, Elon is a software guy. Why is he working on hardware? Why would you, yeah, why would he choose to work on this? Right?

A So you can look at the, like the, cause there's still the, you know, the press of that time is still online. You can just search it. And, and they kept calling me internet guy. Um, so like internet guy, AKA fool is attempting to build a rocket company. Um, so, um, You know, um, that we, we got ridiculed quite a lot. Um, and it, it does sound pretty absurd. Like, internet guy starts Rocket Company. Doesn't sound like a recipe for success, frankly. So I didn't hold it against him. I was like, yeah, you know, admittedly it does sound improbable, and I agree that it's improbable. Um, but fortunately the fourth launch worked, and, um, And, uh, and NASA awarded us, uh, a contract to resupply the space station, uh, and I think that was, like, maybe, I don't know, December, 22nd, or it was like right before Christmas, um, because even the fourth launch working wasn't enough to succeed. You know, NASA also needed, we also needed a big contract to keep us alive, so, um, So I got, I got that call from like the, the, the NASA, NASA team. And I literally, they said, we're, we're awarding you one of the contracts to resupply the space station. I like literally blurted out. I love you guys, which is not normally, you know, what they hear. Um, cause it's usually pretty, you know, sober, but I was like, man, this is a company saver. And then, uh, we, we closed the Tesla financing round. On the …

AI assessment note: “they kept calling me internet guy. Um, so like internet guy, AKA fool”

Q I think you've put this in a really interesting way before, Carrie, where you sort of saying that every founder has become a forward-deployed engineer. That's like a term that traces back to Palantir, and since you were early at Palantir, maybe tell us a little bit about how did forward-deployed engineer become a thing at Palantir, and what can founders learn from it now?

A I mean, I think the whole thesis of Palantir at some level was that, um, if you look at Meta, back then it was called Facebook, or Google, or any of the top software startups, That everyone sort of knew back then. And one of the key recognitions that Peter Thiel and Alex Karp and Stefan Cohen and Joe Lonsdale, Nathan Gettings, like the original founders of Palantir had, was that, uh, go into anywhere in the Fortune 500, go into any government agency in the world, including the United States. And nobody who understands computer science and technology at the level that, you know, at The highest possible level would ever even be in that room. And so Palantir's sort of really, really big idea that they discovered very early was that, uh, the problems that those places face, they're actually multi-billion dollars, sometimes trillion dollar problems. And yet, uh, this was well before AI became a thing, you know, I mean, people were sort of talking about machine learning, but, you know, back then they called it data mining. You know, the world is awash in data. These You know, giant databases of people and things and transactions, and we have no idea what to do with it. That's what Palantir was, is, and still is, that, um, you can go and find the world's best technologists who know how to write software to actually make sense of the world. You know, you have these petabytes of data, a…

AI assessment note: “go into anywhere in the Fortune 500... nobody who understands computer science”

Q startup. Like there's no bottom, there's no double bottom line where people are trying to monetize. Like that's not why it's being funded. So I'm just saying like, if you want people working on this foundational stuff, it seems like if you're trying to replicate what you feel like worked really well, Do you think it needs to be direct? Does the, does the analog need to be directly replicated?

A I'm putting out that it, these kinds of projects don't get funded. Uh, and when you could see it with something like SpaceX and Tesla, um, SpaceX and Tesla both went through major, major funding issues, right? If Elon hadn't been personally wealthy, both projects would have probably failed. Yeah. And, and here, you know, you, you, that is a clear example of something that is, you know, SpaceX might maybe not directly, uh, you know, consumer perspective, but Tesla definitely, right? Like, Tesla is a, is a very consumer-oriented thing, but it was extremely difficult. It was a large-scale, long-term project that just, you know, scared the hell out of VC with good reason. Like, it's extremely unlikely that you would get any of that to work. Um, but what I'm highlighting is not that necessarily VC has to fund this. What I'm saying is that that's not what VC funds. And because that's not what VC funds, and then there is no, you know, no strong ARPA, like, organizing major, large-scale infrastructure endeavors like it used to, then you have this gap and this hole of things that weren't getting funded. And Bell Labs is a great example. Like, I, I, part of the reason that I started Protocol Labs is to try and recreate kind of the spirit of, of, of Bell Labs in an organization, at least focused around, around networks. And, and the only, like, the reason that you had something like Bell …

AI assessment note: “part of the reason that I started Protocol Labs is to try and recreate”

Q to scale as an executive While that company scales, you know, let's look at Sheryl Sandberg as an example. Would Sheryl, if she had never gotten an MBA and never taken the two years out, would she have scaled as an executive and as a leader at Google and then later at Facebook? Or was she enabled or enhanced because of her two years At, you know, the MBA program.

A See, I think I find that tricky, because I think that if you were to grab the average person on the street in the valley, um, and or the average angel investor in the valley, um, They would say that that MBA might have been much more of a filtering and kind of rewarding process than it was an educational process. Um, and I don't know many people in the Valley who would say that you learn more getting an MBA than you learn two years in the grind at an early stage startup. Um, So it's tricky. It's tricky. Cause like, there are a lot of MBAs that have been successful, but there are also a lot of non MBAs that have been extremely successful.

AI assessment note: “that MBA might have been much more of a filtering and kind of rewarding process”

Q How do you develop taste? Yeah. When you don't come from a classically trained world, which would be interesting for next generation.

A Well, you have to, because if you don't, the startup dies, right? So let's say this founder, they go off, they have 95% written by AI. The proof is in a year out, two years out, um, they, you know, have a hundred million users on that thing. You know, does it fall over or not? And then one of the things that's pretty clear is these systems, uh, you know, in the first The first versions of reasoning models , they're not that good at debugging. So you actually would need to descend down into the depths of what's actually happening. And if you can't, then you got, I mean, let's hope that they can go find another architect. They're gonna have to hire someone who can.

AI assessment note: “Well, you have to, because if you don't, the startup dies, right?”

Q know, sometimes it ruffles feathers. Like, do you have any advice for people about an organization and how you. Navigate that really? Like how do you build an org that allows you to think in first principles? Cause if the fortune 500 did that, like the fortune 500 will probably look a lot more like Nvidia than not. And it doesn't like you, you have built a very unique company.

A My state of mind when I'm, my state of mind is always, uh, starts with curiosity. I have a whole bunch of questions myself. And, and, um, of course, like anybody else, I'll seek the shortest path to the answer, um, but oftentimes the, the answers from the people that are near me, uh, might not be satisfying, and, and I might have other questions, and maybe they're, they're busy doing something, and they're pursuing something, and so my first, my first inclination is to go discover the answers to my own curiosity. Um, my second is if I find that the information is and that the domain of information or, you know, particular field, uh, could be really important to somebody and could be important to our company, then my next inclination is how can I learn as much as possible so that I could be of service to the company and share with everybody else? You know, this is no different than, than you when you're, you're sharing knowledge. I mean, I watch your podcasts and I watch your, your videos and I really enjoy them. You're sharing ideas with everybody else. In a lot of ways, I think a, a CEO is in service of the company, in service of all the people that are working there, and you want to empower them with some insight. And so that's really where it's coming from. It's not so much a management technique, but a personality technique. You know, I, I want to empower you, and this is s…

AI assessment note: “It's not so much a management technique, but a personality technique.”

Q Now I guess one curious, uh, now shifting gears a little bit about what are your thoughts in terms of how the future is going to look with, uh, coding?

A We were kind of this, this, Maybe middle road bet from the start, where when we set out to work on the company and we were kind of hiring our, our first people, we would get these weird looks around, you know, why are you, I mean, at the end of 2022, it wasn't really like this, right, because kind of ChatGPT happened and then the whole world woke up to things, you know, beginning of 23, but especially during 2022, uh, when we were working on the CAD stuff and then the early code stuff, um, people thought working on AI, uh, was, it was kind of weird to do. People were not entirely convinced that it was a good use of time and that there were going to be lots of great applications to follow out of AI. And then even the people who are interested in AI There was, I think, in our space, you know, a bunch of people that were just focused on optimizing kind of the form factor that exists already, um, and just making those products a little bit better. And then at the same time, you know, in our social circles and professional circles, there's a bunch of people that, you know, were thinking, oh, why would you work on anything other than AGI? And, you know, all of the work that you're doing right now in one or two years, you know, circa in 20, 22 is going to go away. And Yeah, I think that we've always had this view that there's going to be lots and lots of, um, incredibly valuable thing…

AI assessment note: “we've always had this view that there's going to be lots and lots”

Q of the world, the, you know, the corporate world, like the world at large that does not understand what's happening with AI, they're going to look to the people in this room, uh, for exactly that. It sounds like, well, you know, what are some of the tangible lessons? It sounds like one of them is don't give up board control or be careful about have a really good lawyer.

A Uh, I guess for the first, my first startup, the, the big, the, the, really the mistake was having too much, uh, shareholder and board control from legacy media companies who then necessarily see things through the lens of legacy media and, uh, that they'll kind of make you do things that seem sensible to them, but, but aren't, really don't make sense with the new technology. Um, I, I know, I should point out that I, that I, um, I didn't actually at first intend to start a company. I, like I tried to get a job at Netscape. Um, I sent my resume into Netscape and Mark Hendrickson knows about this. Um, and, uh, but I don't think he ever saw my resume and then nobody responded. So, uh, and then I tried hanging out in the lobby of Netscape to see if I could like bump into someone, but I was like too shy to talk and talk to anyone. So I'm like, man, this is ridiculous. So I'll just write stuff for myself and see how it goes. So it wasn't actually from the standpoint of like, I want to start a company. I just want to be part of building, you know, the internet, uh, in some way. Um, and, um, and since I couldn't get a job at an internet company, I had to start an internet company. Anyway, the, yeah, yeah. I mean, from an, AI will so profoundly change the future. It's difficult to fathom, um, how much, but, You know, the economy, assuming we don't, things don't go awry, and like AI does…

AI assessment note: “mistake was having too much, uh, shareholder and board control from legacy media companies”