The Exchanges

Q Can you tell me some stories about like success stories from Replit? Maybe like the, the, what's possible as a solo dev today? I mean, we're seeing Peter levels IO do crazy stuff, but what are you seeing from the, from the most, uh, successful Replit users?

A Let me start with the, with the business cases. Cause I think those are, are kind of really profound. Cause we've always had like micro entrepreneurs like make money and you know, it's certainly accelerated massively, but A hundred plus year old, year old company, uh, Sears. Um, I don't know about you. I didn't know they, they continued to exist, but they, they have this, um, uh, spinoff called Sears Home Services. About a year ago, uh, they hired a team to Like a more trendy, uh, sort of, uh, tech team to get them off of Cobalt, which is whatever a hundred year old program language. And they were able to do it very quickly. And then after they'd done that, you know, basically most, most companies as they're modernizing, they go adopt a bunch of SAS tools and no code tools and local tools and to become like a modern company. They actually skipped all of that and went to, went to AI and they're Operations team, which manages the other kind of field workers that go out and every day kind of do house repairs and all of that. Uh, they start, they started building AI tools. So one AI tool that they have is like the worker would, would go and would talk to the AI and it would optimize it at their route and would, you know, that made it so that they're making a lot more money. And now that team, everyone in that team started building AI tools. And so you have this ops theme that's hig…

AI assessment note: “they have this, um, spinoff called Sears Home Services.”

Q you know, we, we've seen plenty of, you know, undersea cables being cut, uh, How worried is the average space founder generally, uh, like how worried are they about sort of sabotage? Is that something that's, that's sort of like a risk factor in a, in a series C deck where they're, where they're like, you know, uh, you know, is that, is that even the right question to ask?

A I would say like not too worried. And it's not to say that it could happen or anything and we should underwrite this, but it's like the, the, the thing that would be really bad is if like Russia basically, you know, nuked, uh, uh, low earth orbit, you know, detonated through nuclear weapon. And you had a bunch of like highly charged ion particles, you know, basically orbiting around, uh, earth at a much lower level than where you like the Van Allen belt is. Cause it would basically destroy everybody's Leo business right now. Part of why Starlink works is they're low enough. Um, in orbit that, uh, the earth's magnetosphere basically deflects the solar winds. Um, so you're, you're still relatively protected. If that all of a sudden becomes a very unprotected zone, you have a ton of radiation, man, everybody's economics basically just blow up. Actually, one of the few that probably doesn't blow up is BARDA because radiation basically screws you on a day by day basis. And we don't need to be up there for very long, but anybody who's business case is built off of like being in space for a longer period of time. Man, if the Russians launch a nuke up there, that'd be really hard. And I haven't seen any, like, how one would fix that problem if they did. Like, I don't, I don't think that, like, there's even, like, a hypothesis for how to do that. Like, you know, maybe some, like, really…

AI assessment note: “I would say like not too worried.”

Q So right now you have the subscription. What, what is synthesis look like longterm starting digital makes a lot of sense, but over time is there, you know, desire to kind of almost verticalize and go into the real world or what does that look like?

A I don't know about going into the real world. I think it's, there've been some, You know, Silicon Valley startups have been really well funded that tried to do the software piece and the real world piece at the same time. I think it's really hard to pull off both. Like if you're going into the real world for schools, you're doing like a real estate play essentially. Um, and, and, you know, I think when it comes to learning, you know, math, science, engineering, the fundamentals, these are, this is information. And so information can be transmitted, uh, through computers. Um, In a lot of ways, you know, better than, than a human can do it. And so I think if we want to reach a billion kids, then, you know, it makes sense to go vertical, but within, you know, software, like things that can be delivered purely software, never say never, because it is attractive to, uh, you know, the idea of like opening up physical in-person schools. If we did something like that, we'd probably do something like an academy. Uh, like the, I always admire that the European soccer clubs, what they do is they'll have an academy attached to all their professional teams. And they'll bring kids in at, like, age six or eight. There's a guy who's a star now for Barcelona that they brought in at eight years old, and now he's, like, 18, and he's just, he's amazing. I think if you really want, like, incredible…

AI assessment note: “I don't know about going into the real world. I think it's really hard”

Q kind of around you, and whenever there'd be a profile, Isaiah would be there, a couple other guys would be there, the nearest crew, the Pico grid crew, a couple other folks, Dirac, right? Uh, but is there, is there a new addition that we need an update on Gundo, or is there just like a whole separate crew that you're not even overseeing as the godfather of the Gundo?

A Dude, yeah, so the Gundo book club is a lot of these dudes, rather than being like, I don't know, 23 through 27 They're like, 19 through 23. Totally different cohort. A lot of cool companies. Teddy cohort. Durin finally made it, uh, to the Gundo from Georgia. That's really cool. Aries is sick. Um, uh, Chris, what was his last name? Ottoman. Um, yeah, he is building subsea vehicles. There's a whole new click and, uh, we're almost out of touch with us, you know, part of what made Gundo possible and like the initial vibe real possible was we were all running like pre-seed companies. So we had a lot of time on our hands and kind of just like around. And so we didn't even get to know these guys as well. Um, but they're a super tight click, like the original founding Group of gundo bros.

AI assessment note: “There's a whole new click and, uh, we're almost out of touch”

Q That's sort of helpful, right? It's like, it's actually great that the, uh, uh, uh, you know, attention is going there, uh, and, and there's plenty of opportunity and sort of low earth orbit for the rest of us.

A Yeah. So I think that's a great point. Um, Look, I think, I think SpaceX is deeply incentivized to have a giant ecosystem of lots of great independent companies in space, and I think especially in LEO. And yeah, you're right. I mean, I think the focus is going to be Mars in addition to Starlink, which is already, you know, operating and scaling. And on Mars, I'll just say, I was I mean, I've always thought just from an intellectual perspective, it's the coolest thing ever to build a city on Mars. But like when I wrote my initial SpaceX investment memo, you know, it was like, yeah, I don't really understand Mars from a business perspective, but, but like, I think this is going to be such a great company. Yeah. Like irrespective of it, that it's fine. Like it's a side quest from a business perspective. I am now at the point where I actually, I really believe that Mars is going to Be like one of the biggest profit generating endeavors in history, and not necessarily, not like from the direct Like building a city on Mars, but from all the indirect things, and I'm sure you guys remember all the myths about, or, you know, all the stories about NASA in the sixties and seventies, all the spin-off technology, microwaves, materials, and all these different things. Um, I think the spin-off technologies from having a clear goal of building a self-sustaining city on Mars, which will have ab…

AI assessment note: “SpaceX is deeply incentivized to have a giant ecosystem of lots of great independent companies”

Q Has any, has anyone at Anthropic, has anyone at Anthropic commented on this at all?

A I haven't seen anything. There is news out of, uh, out of Google. Uh, a Google paper warns that, uh, warns crypto on quantum risk ahead of 2029 timeline. So we've heard about the risk of quantum computing affecting Uh, the cryptocurrency industry, crypto projects broadly, uh, there is some new, uh, research out of Google that provides some more perspective. So, Google researchers have warned that future quantum computers may be able to break some of the cryptography protecting Bitcoin and other digital assets with fewer resources than previously thought, adding urgency to the debate over how the industry should prepare. The researchers did not indicate such a machine exists today, But said new work suggests the computing power needed to carry out that kind of attack may be lower than earlier estimates had suggested. In a Google research blog post, this is from Bloomberg, uh, the researcher said that a future quantum computer could break elliptic curve cryptography, a form of public key encryption used across much of the market. Their latest estimate points to a twenty-fold reduction in the quantum computing hardware needed to break what's known as ECDLP- A mathematical problem that helps secure crypto wallets in transactions. That does not mean Bitcoin and Ethereum are suddenly exposed, but the researchers in the white paper dated Monday said the clearest defense is a shift tow…

AI assessment note: “I haven't seen anything. There is news out of, uh, out of Google.”

Q Like, why, why is it such an incredible opportunity?

A We have bought more than a hundred companies. You're 100% right. Over 26 years. We've bought more than a hundred companies. We've invested in hundreds of companies. We own one percent of Anthropic. You know, we just sold a company to Google, which is Wiz, the security company. We, you know, we grew Snowflake and took it public. We're investors in many companies. Um, so we want to talk to you if you're an entrepreneur. We want to invest in you. We want to help you grow. We want to bring you here, introduce you to all of our customers. We have 50,000 customers here to introduce you to who want to buy your product now. So come here to San Francisco or come to one of our world tours all over the world. You know, from here we go on the road and we're going to be in every country in the world, every major city, and we will come and meet you. We want to meet you. We want to know you. Salesforce ventures is our arm that acquires, but also invests. It's a huge part of our business. It's a five billion dollar fund. It's I think doing a 33% IRR. I don't know exactly the numbers, probably one of the highest performing funds in the world.

AI assessment note: “We have 50,000 customers here to introduce you to who want to buy your product”

Q 40 different accounts that are really all dedicated to Um, clearly content and they're just spamming, spamming, spamming, spamming, spamming, just like shotgun approach seeing they'll put out a hundred videos. If a lot of them get under a thousand views and spam one gets a million, that's, that's a win for them. Um, and, uh, yeah, just, I don't know how long this will be like the meta, right?

A Well, so it's like, I think mediocre work is like the super ripe for AI automation. And we're already seeing that where I believe Substack, we were talking to Chris Bast, and Substack will automatically generate clips. There's another company called Opus Clip. We've built something internally, Newsmax. That does some automated clipping, but actually telling a story around a piece of content that's within a larger piece is something that AI is not able- Clipping intelligence has not been achieved internally. Yeah, because you, in order to do something that doesn't just outperform on views, but also outperforms on brand and holds the brand standard, it does somewhat require like human ingenuity. And we saw this with, uh, we talked to Dwarkesh about this. Where he has spent a lot of time using anthropic and Gemini and open AI to create workflows that will go through the transcript and try and find the best clips that will perform. And he says it's always comes back at a five out of 10. And so he still has a team that does this and is very good and has a whole bunch of like hard one lessons about what works at a certain amount of time. And we've seen this with our, with our content that actually winds up performing a lot of it. Like we, we, we have to do a lot of experimentation. Um, but once we do the experimentation, we usually learn a lesson from it and then we commit that to me…

AI assessment note: “mediocre work is like the super ripe for AI automation. And we're already seeing that”

Q Uh, can you talk about the extortion attack, how you handled it? And I'm particularly interested in going forward. There's a lot of tools in your toolkit. You can throw more money, more people, more technology at the problem. What's gonna, what's the next few years look like in terms of beefing up security?

A I mean, we've had a really good track record on security at Coinbase over the last 12 years. And even in this attack, you know, we didn't see any private keys or, Funds access directly by the attacker. So what they, unfortunately, they were able to do is they were able to bribe some of our customer support agents and to share with them, like, personal information on, on customers, like name, address, etc. And, you know, they want this information. Attackers want this information because they want to text people, call people, try to socially engineer them, and to get them to send their money. And so, unfortunately, they were able to successfully do that with a handful of customers, which we've reimbursed a hundred percent at this point. Um, and then they sent us this demand once, once we were, you know, they realized they'd been found out, and customers were kind of getting educated about this, um, they sent us this demand for twenty million dollars, or we're gonna release all this information. And it was an interesting moment, you know, we came together and thought about it, and, um, you know, on the one hand, we could've paid them the money, right? But we were, we, we talked to a bunch of security researchers and people, and they were like, you know, there's really, What typically happens in these cases is you, if you pay the extortion, they're just going to come back in three…

AI assessment note: “we put out our twenty million dollar bounty for any information”

Q I actually think in the era of just AI slot products, these, like, ultra, ultra handmade, like, it will become cool again for somebody to say, Yep. I'm gonna hand make wine bottles, right?

A And, and this is why you can't write this off as just like, oh, it's just rich guy shit, because Zuck is a technologist, and watchmaking is in some way a pinnacle of technology. Like we didn't have the ability to make this. It's just mechanical. It's not computer science, but there's something that's when you look at a really fine handmade watch, you're, you're realizing like an expertise in craftsmanship that you want to bring to. Yes. Even the ad platform software like that. That's the same, same concept. Um, the dial and finishing the minute track. And seconds subdial are executed in true-fired enamel, requiring multiple firings at 800 degrees Celsius. The silver bridges, known as malacourt in French, they feature a frosted finish achieved by repetitive dabbing with a stiff wire brush. The copper content gives these bridges their soft golden hue. Mirror polishing, every bridge and wheel includes expertly polished bevels. Many Feature sharp inward angles, which are extremely challenging to master. Each screw goes through 12 different manual steps and can require around eight hours to produce. Finishes vary between black polish and fired blue steel, all color matched meticulously. Incredible. A one-minute tourbillon made of 69 parts. The tourbillon, I gotta get a tourbillon watch for sure. It's so sick. Originally, I believe the tourbillon was designed to Offset the, uh, I thi…

AI assessment note: “when you look at a really fine handmade watch, you're, you're realizing like an expertise in craftsmanship”

Q Uh, talk about the, the, the fans. Are there really Elon Musk fans in the courtroom? Like what, what motivates someone to go and watch that live? Is this their UFC front row ticket? Like why are they there?

A Uh, so you guys would have fun. Like, it actually is, a lot of court cases are boring to people who don't care about this stuff, right? Like, you and I may be super into, like, the FTC trial was super fun for me because it's like, oh my god, Mark Zuckerberg emailing Sheryl Sandberg and talking about PATH. Like, this is incredible. And like, the average person has no idea what we're talking about. But this is like a circus. There are people who genuinely love Elon, Or are genuinely worried about the end of the world happening. Um, and I think it's a really good thing that there's public access to these courts. Like, I think, like, the average person can come in and show up, and that's what, I think after the, the buzz of Twitter and, like, people seeing that this is an event, we got a much longer lines, and, like, folks who are local, like, just, like, I know a PM in tech from Meta that came, I know, like, a guy from Box made it in, Uh, just, if you get a seat, if you get there early enough, then you can get a seat, and you can just hang out. It's like, and I think it's really great. I think it's great that people are, yeah, are there for it, you know. It shouldn't just be me.

AI assessment note: “There are people who genuinely love Elon, Or are genuinely worried about the end”

Q You mentioned that the space was crowded. How did you think about finding differentiation on the product side, the distribution side? Uh, what really stuck out is like, okay, it is a complicated space, but there's a big opportunity here.

A I, I would say, man, you know, you, you have to have access. So Juan Domingo Beckman Jr. Uh, of course the family, what they've done in this space, you know, you're talking about generations on generations of success and growth, um, in the business, right? Like from a distribution, you know, uh, outlet or opportunity, there isn't a bigger option. So, uh, I think for us, it was getting Juan to understand the real want Um, getting him to understand the passion behind my work and that this is not a celebrity play. This is not a, a check grab and run. This is a, want to build something that can literally be attached to my family name and give me an opportunity to, to build generational success, uh, wealth, uh, visibility, whatever you want to call it. But I want, I want that. How do I get that? You have to go to the people that have it. So, um, I think that was our biggest, I'm biggest one. And, and as you talk about separating yourself, well, when you have that machine and that machine understands your real energy and those two things connect, well, you're already so much different from anything else in the space, right? Like the space is crowded because people believe that, oh, let's get a famous person and put the famous person name on a bottle and just put the bottle on the shelf and it's yourself. That's not true at all. Like, why did you make it? What's the story behind it? W…

AI assessment note: “from a distribution, you know, uh, outlet or opportunity, there isn't a bigger option”

Q Was it, was it just about open source, or was it about sharing research?

A It was, it was, he was saying that you don't have a, you don't have a, uh, an obligation to open source the weights of a model that cost a billion dollars to train. He's not arguing for that, but he is saying that you should open source the tricks, the, the research ideas. Basically, publish the research papers again, and empower a broader community. Of course, there's a lot of, uh, There's a lot of competitive dynamics there, but that is something that could potentially, uh, happen via regulation or something or happen just, just due to just a competitive dynamic. Like there are other labs out there that don't have as much compute and might realize that they have great researchers and maybe they want to open source more. Um, there's, there are a lot of different ways that this could play out. So, uh, he says, I do not think it's right that a few labs, few AI labs would make the most consequential decisions About the shape of our future. And so there's been a bunch of, a bunch of back and forth about the attacks and what's driving them and how, uh, how risky the rhetoric has been. I think in general, it's, um, it's a very tough situation because the, you don't want to just spark more controversy and more discussion around this stuff. You mostly want to move towards more security and more, um, more.

AI assessment note: “he is saying that you should open source the tricks, the, the research ideas.”

Q so what do you mean, if Accenture folks are using Cloud Code, wouldn't they be using it on consulting projects internally to companies?

A Yes. I, I, I think they're going to be using it internally, uh, as much, no, so they're going to be using it internally, and then they're going to be doing all these consulting things, because if you think about it, one of the issues is like, You know, when you had SaaS, one of the biggest issues of like changing from one CRM to another was effectively being like, Hey, everyone, you're gonna have to quit your jobs for like 10 months to figure this out. Um, the implementation, the implementation, like you'd have tons and tons of like consultants do that. And I think that that's what, uh, the Accenture partnership is. So essentially like people are going to be implementing cloud code and there's 30,000 people at Accenture who's going to do it. And then on the other side, you have Frontier, which is like the fortune 500 way of being like, Here's your plan. Come to us and we'll build this whole thing. Blah, blah, blah. So yeah.

AI assessment note: “Yes. I, I, I think they're going to be using it internally”

Q Yeah, what's the, uh, what, what's, what's the update on that front?

A I mean, we're, we're knocking on the door. We might be able to get in the playoffs here. Um, it's gonna be, it's a little tough, a little tough right now, but I'm hoping, I'm hoping we get in there, but that, that was probably one of my favorite, like, I was like March of 19 when I rage tweeted about how undervalued women's professional sports was and in, in proper Twitter fashion, most people said I was an idiot. No one cared about women's sports, but that, that obviously has done pretty well. And, uh, and yeah, thank him for being an angel city supporter. I've got a couple extra spots in the suite. If you've got, there's a couple of listeners who want, uh, we'll make it happen. You have to figure out a way to give it away here for the last home match of the season, but I'll gladly.

AI assessment note: “we're knocking on the door. We might be able to get in the playoffs”

Q You have been extremely busy. Give us the update. What's new in your world?

A Gosh, what's new on our world? Well, the, the big news is we, we launched dynamic tech defense reform, which I know we're going to get into today, um, which is a huge initiative on behalf of all of defense tech, all of the defense tech community. Um, but also it's, I just returned from Washington last week, had a host of conversations across the Pentagon, across Congress, and it feels like for the first time in probably 10 years since I've been doing this, that everyone is aligned, that we need to see Major, major changes happen to defense procurement, and not only is that sort of things that people talk about, but it feels like it's finally going to happen in this year's NDAA, um, and that we're on the one yard line, um, sprinting to a finish that I think will, will hopefully, um, end up great for, for all of the, the companies in the defense tech community, so we can go further into what's changing, but I always, like, I, I feel like I'm bringing, like, actual good policy news, like, probably, like, It's like the one thing that everyone in Washington can agree on, Republican or Democrat, uh, you know, House or Senate, Pentagon, everyone is aligned.

AI assessment note: “the big news is we, we launched dynamic tech defense reform”

Q kind of like the big moments that you're kind of tracking the, the open AI open source model could be one, whatever meta launches next, right? I'm assuming they're going to be dark until this new team can, can really cook and bring something great. Maybe they don't do anything this year, but I would assume they, they come with, with something. What else are you, are you tracking? Yeah.

A I mean, for me, the O three release in chat CPT was like a pretty like game changer kind of thing where it was like, we saw with like deep research that like, okay, they kind of figured out how to make agents work, but it was also just like this one version of an agent and O three, you can kind of get it to be a pretty general agent where it can like do some pretty complex stuff that was kind of new to see from like the geo guesser thing was crazy. Um, And having that as a like vision of like what AGI starts to look like, I think it's pretty cool. Of course, from the like research open source world, there was a deep seek as like the RL craze taking off, but like, I mean, I'm, I work on RL, so I like obsess over it and like think about it a lot, but I do think we're really starting to see these recipes, at least in the broad strokes of like, okay, here's how the LLM thing can go. We figure out what we want it to do. We give it some tools. We set up these environments. We figure out how to evaluate it. And then we can just kind of like let it go. And these things get better at doing those things via trial and error. Um, And so like, I like, I think that is one way to kind of forecast where things are going. It's just like, what are the plausible use cases that people want to use elements for? They want an agent to do X, Y, Z. Um, and then how do you make this a thing that you can…

AI assessment note: “for me, the O three release in chat CPT was like a pretty like game changer”

Q No, it's very economically rational. Uh, what about the, uh, conjecture that, that you had, uh, a, a, a team of junior developers underneath you helping you actually accomplish the tasks? Is there any truth to that?

A I wish I had the money and I wish that was true, but that is not true. Um, any of the founders that, uh, you know, uh, I worked with can watch with that. I have multiple locations like pair program with people, you know, I've written every single inch of the court. Uh, you know, anti-metals, one of the companies I worked at, uh, you know, they have pretty high bar for design, as you might know, uh, they did a launch recently. Um, you know, there were two occasions where, you know, just before the launch, like, You know, the, the contractor who was going to do the front end for us essentially had bailed out and, you know, we as a team had to figure out and do everything, which was a team of only three engineers, right?

AI assessment note: “I wish that was true, but that is not true.”

Q So was this, was this, uh, is this fake or is this real?

A I think this is a brightness adjusted. I think that they, I think that whoever shared this image, uh, turned up the brightness or, or brought up the, brought up the lows. Uh, D basically decrease the contrast in the, in the photos app, like, uh, to make it look worse than, than it actually would. Uh, but I mean, even, it doesn't even matter because even if this is the way it looks, this isn't the way it will look when it will ship to, to, to people because it's obvious and, and, and, and Apple doesn't make these like obvious mistakes like this, especially when it comes to design. Um, they, they, they might do something that is like rough in the short term, but better in the long term, like what I think is happening with the photos app where In the photos app, they are understanding that in the future, you won't want to scroll at all. You will just search and search will be so powerful because every photo and video will be tagged so thoroughly from AI and it will understand the context so much that you will just ask it to pull up a photo and it will do it immediately. The problem is, is that right now people, a aren't, they haven't switched that UX paradigm to actually go search first. And second, um, And, and, and second, just the tagging is not that good. Like I had a video clip of Arnold Schwarzenegger from a movie on my phone. I wanted to pull it up. And so I searched like A…

AI assessment note: “I think this is a brightness adjusted.”

Q Well, anything in voice agents, or just improving the customer experience broadly, like surprise and delight, right?

A Well, you know, it's one of those things that I think is funny, like people talk about hallucinations as this bad thing, but in a lot of ways, for like, kind of this speech-to-speech interaction model, where you're talking to an AI, Hallucination is a feature. So if you remember like your Alexa or Siri, like when you were talking about this, you know, 20, 10, 20, 12, you go and you ask it a question and it's just going through a bunch of like a decision tree. And if it hits, you know, a dead end and it can't answer the question, it just doesn't do anything or it says I can't answer the question. And so the, the second that it, it can't do something you expect it to do all of a sudden you, you know, it's a, it's such a punishing user experience. You just don't want to use the thing anymore and you stop using the thing, but. With kind of these new models that actually, you know, hallucinate answers into existence, even if it doesn't know the answer, it tries, um, you always get a response, uh, that comes back from the model. And for, for like voice interfaces that actually can be more of a feature than a bug, um, it also helps that these models.

AI assessment note: “for like, kind of this speech-to-speech interaction model... Hallucination is a feature”

Q returns, it might judge me, and if I'm living an immoral life, the AI God would sentence me to essentially like an AI version of hell. It's not a problem to live an immoral life in the absence of God, but if God returns in the form of this AI God, then there will be a reconciliation moment. Is, do I need to put on the tinfoil hat for that?

A Well, I mean, I personally think that, uh, oh, there we go. There we go. Put it on, baby. I mean, personally, I think that, uh, if you read the Tower of Babel story, I think that, uh, a lot of my faith is, I don't think these AIs can become gods. I think that something will happen, and we see that in the Tower of Babel story. You know, you've seen people come out and say, hey, we're building these new gods. Um, read Psalm Humans try to create gods. We've seen this story play out, and to come back to some of your earlier questions about tech and utilitarianism, look, a lot of the peace that I have with what's happening with AI, I'd be freaking out if it wasn't for my faith. I mean, I have complete faith that God is in charge. He'll take care of this, and when I read books like the Book of Judges, I see people turn away from God all the time. When I read the Old Testament, the Exodus, you know, you know, you see the golden calf, and I just see AI as another version of that.

AI assessment note: “Put it on, baby. I mean, personally, I think that... I don't think these AIs can become gods.”

Q guys' sort of net burn. Uh, it's really astonishing, right? And it's gotta be, if you're running a business, you know, at, at anywhere in the same category, even adjacent categories, you have to be highly concerned at the, at the rate that you guys are growing in that context. Uh, is that how much of that do you credit to ramp versus yourself and your team's, uh, operational abilities?

A I think our whole mission, right, to your point is, like, make the most of every dollar and every hour. If you're spending it, you should know why, um, uh, and it's not just that you're spending it in one place, but, but what actually generates the return? What, what, what connects to growth? Um, and I, I, I would say that, um, we certainly benefit a lot from our own product. Um, uh, we try to be obsessive over making sure we're not wasting, and if we're spending on, on, Um, you know, software license is a product. We buy 500 seats. We want 500 people to use it. And we use RAMP to cut out extra licenses. Um, we look and make sure we're getting the, the most impact for, for prices and, um, you know, software that we're buying. And, uh, I, I just think, you know, for the same reason, a lot of the highest performing best running companies, um, on the planet run on RAMP and use it. Um, you know, we do too. We, we want to be the best customer of our own product. And so some of it's that, um, but I, I'd also say with your whole mission is, You know, to obsess over getting, you know, your customers the most from every dollar an hour. Um, as long as you're not a big hypocrite, you start thinking about it for yourself too. And so, um, I, I would give credit to the team. Uh, I just think that the rigor, um, you know, not just financially, but even just from a, you know, engineering persp…

AI assessment note: “some of it's that, um, but I, I'd also say... I would give credit to the team”

Q one hundred and eighty seven billion in line with analyst expectations. Net income rose 88% to twenty billion higher than predicted. So very good year last year, a little bit rougher going forward. Revenue from Amazon's cloud computing unit, AWS, increased 19% to 28.79 billion, slightly lower than expected. Analysts said they were closely monitoring this number after disappointing cloud computing results from Alphabet and Microsoft. What you thinking, Jordy?

A Uh, I'm thinking that Amazon has been weirdly out of the headlines on all things AI CapEx. It's just not something that you're thinking about and not really hearing about much. I mean, obviously people like Dylan Patel are spending a lot of time writing about it, but it hasn't been top of mind. For a lot of people, but given how, how much of their business is in cloud, you would think that they would be using that to drive the narrative and trying to build momentum around, um, you know, the narrative, the Amazon narrative around AI. And it just doesn't feel like. It doesn't feel like, you know, they obviously have like the actual, like core data center infrastructure and cloud leadership, but it doesn't feel like there's no, um, They don't, they don't have their like sought yet, right? Like somebody who's coming out and, you know, going on CNBC, trying to get on technology brothers, likely, uh, things, things like that. So they just don't feel like they're, um, while they are like critical and they're running a lot of this, you know, work, they don't feel like they're a big part of the conversation right now.

AI assessment note: “I'm thinking that Amazon has been weirdly out of the headlines on all things AI”

Q that there's a trade-off there if you go with something, uh, more produced that's not just you opening your laptop and yapping, right? Uh, You can get stuck in like the production quagmire if you're trying to build the Mona Lisa or some incredible show. At the same time, there's prestige that comes with that. There's a new level of authority. Uh, how do you think about that trade off?

A I think it's barbell. So it's either like make it the Mona Lisa and like play to win that game, or you're in the volume game and the volume is just value per second and just trying to get as many of those seconds out as you possibly can. And like, I just think it's, it's, it's both. And They're, I think they serve, they serve different objectives to your point of like you gain authority in different ways. It's definitely, there's an element that it shifts the brand. There's access to new distribution and eyeballs. All of that happens over here, but over here I think is where you get a lot of the closer to buying behavior, believe it or not. And so I see this as more top of funnel and this is like closer to middle and bottom. Obviously you can get discoverability. Like I said, this is full stack, but you get way closer to purchase it on the side.

AI assessment note: “I think it's barbell. So it's either like make it the Mona Lisa”

Q Uh, what, what, what's next for you at Meta? What, what, what are you focused on? What do you want to deliver this year?

A Yeah, for us, there's, look, the really big, big arcs, obviously are continued towards full AR, so we're really excited that we're supporting the entire strata we talked about before. Full AR, display glasses, regular glasses, and, and even other, other exciting fun factors that I can't yet tease. Um, on the VR side, we're advancing the hardware. We have multiple fronts that we're advancing the hardware on, and also on the software side, supporting creators. Gen AI. Don't sleep on it. That is the real unlock. You've got Roblox, you've got Minecraft, they're awesome tools, you've got tremendous communities there of creators, but there is a ceiling on how good the rendering can be in those platforms, which I know, and it's also, you have to be a certain level of creator to be able to produce good content in those platforms. Um, and with Gen.ai, you can lower the floor, so it's so much to do.

AI assessment note: “the really big, big arcs, obviously are continued towards full AR”

Q another, uh, topic that I'm, I'm sure you guys, uh, so one, uh, I don't think AI is going to take your guys' jobs anytime soon, uh, despite it being able to maybe piece together, you know, random ideas, uh, so much about, uh, taste and things like that. But, uh, how excited are you to leverage that technology for, for some of, um, for some of these various plays?

A Honestly, TBD. And I think, I think it's, I do, I will say like, unlike the whole, like crypto craziness of the last couple of years, um, we, we didn't get into that cause we didn't feel like it was here to stay yet. That that's like the truth. Like when you see the herd run so fast, it usually means they're running towards the end of a cliff. There's a little bit of a similar sentiment right now that said, I, it is rooted in something real. Um, but we're not in a rush to like. Use it necessarily, right? If it's going to be here forever, we've got time. Like it's mischief doesn't win as a first mover. If you look at all of our work, it was never about being the first to react to something or the first to like, make use of a format. Um, so we'll see. We just have to see.

AI assessment note: “Honestly, TBD... we're not in a rush to like. Use it necessarily”

Q What's the whole story with packs too? It was like the same company.

A Yep. Yep. So when they started, they were doing, uh, What was it? They did, they had flume or flume or something. I forget what it was. They had, they had a different tobacco vaporizer, and then they had, and then they had the same kind of technology to heat up material, and you could, in theory, put tobacco leaves in there, and then vaporize that, and just warm it up, and breathe that in, and that would, it would be like smoking a pipe, or like vaping, Tobacco, I guess, loosely, but obviously like everyone was just using it for cannabis. And so that takes off, that becomes this like fantastic business on its own. Um, and then they had this other product, I forget what it was called, but it was a, it was a tobacco vaporizer, like an e-cigarette, similar to Juul. They wound up selling that to JTI in Japan, or like doing this crazy licensing deal to get that out. Then they wind up splitting the company once both products are kind of taking off, but they're on very, very different trajectories. And it's very clear that They will be under very different regulatory regimes because the FDA is set up in a way that there are a number of different organizations within the FDA. So the FDA approves cancer drugs, and that's in the FDA drugs. They approve biologics. They approve veterinary medicine. They approve, uh, medical devices. So like when you go in an MRI machine, the FDA has approv…

AI assessment note: “Then they wind up splitting the company once both products are kind of taking off”

Q Wait, what is this photo? That's not, what is that? The team just accidentally put in some random car.

A Okay, so, so the car, the car is strange. I love the interior, but the car overall is strange. It's still unclear to me who it's really for, but they are finding buyers. The idea that, that Buying the Luce, like, Ferrari is saying we're not forcing any dealer to push this car or whatever, but the idea that it's not going to have some, whether or not you bought a Luce is going to have some weight on your future allocations to me is just insane. There's just no way that that's true because every single dealer is going to look at their client list. They're going to look what cars have they purchased. Like, we know multiple Ferrari collectors that are buying Two separate cars that they don't even want of the same style in order to gain status within the dealership and show that they're a proper, proper collector and they're properly sort of cherishing the brand.

AI assessment note: “the car is strange. I love the interior, but the car overall is strange.”

Q But it is, it has been fascinating to see, like, does this actually need to be an app?

A Yeah. Yeah. I mean, there's, there, there's a ton of apps and software that will still be valuable, whether it has a liquidity pool or some sort of unique, uh, source of strength or some differentiation point that, that the existing chat apps can't hack at all. And then there's also just like, like marketing ARBs effectively where it's like, okay, yes, the, any, any frontier model in a chat app could do this, but you weren't aware of it. And this company was really good at running. Ads to actually get awareness going and then drive downloads of this specific event. And so we see those in the, uh, uh, in the app store all the time. The other thing that I was, uh, uh, reminded of was, did you ever read Union Square Ventures, 2016 blog post fat protocols? Are you familiar with this? So fat protocols, uh, was this concept around how in, in the web, like web, I guess, one point oh, two point oh, there were protocol layers, which are like TCP, IP, HTTP, SMTP, like file transfer protocols, HTTP. And there, and a lot for like a couple of years, the crypto community was like the group that developed and maintained HTTP. They basically created the standard that the web ran on and yet very little value accrued to The creators and the maintainers of that protocol and crypto would be different because the Bitcoin protocol had the value capture component like baked into it. And so there was …

AI assessment note: “there's a ton of apps and software that will still be valuable”

Q Which is like, I mean, he's 54, he's got a while, right?

A Yeah. If he's a radical life extension guy, he could be like, 300 years from now. Who knows? Uh, that's a very vague timeline. Where, whereas in the past with Point to Point, with Roadster, with Cybertruck, with Space Data Centers, all the other Elon pitches have been very focused on, uh, we think we can do this in five years. And he always gets criticized for being Elon math, too aggressive, but then history's shown that, you know, he, he, Over promises, under delivers, but he typically does deliver eventually. And, and if the technology is completely new, no one else was going to build it, and he's the first one to build it, it's fine, because Starlink works. It was the first Leo constellation for satellite internet, had turned into a huge business, wound up being a great technology. And so, uh, now the mass driver's a lot crazier. I was, uh, going back and forth with Tyler about this, trying to get him to, uh, nail down a prediction. We started with 100 years. Do you think it's possible to put a mass driver on the moon in the, in the next 100 years? And you said yes. So it's possible.

AI assessment note: “If he's a radical life extension guy, he could be like, 300 years”