Latest / Newcomer Pod / The Casino enters the arena + Guillermo Rauch of Vercel on AI factory builders
Transcript
- 0:00I think we need to break down the logic of why this week
- 0:03people decided that the AI bubble has burst.
- 0:05And there were some key things that happened that we, I think
- 0:07we can deconstruct and decide whether or not they're actually
- 0:10worth, you know, being a hater over.
- 0:12The first one was Sam Altman had a dinner that I was not invited
- 0:15to with reporters where he discussed, among other things,
- 0:19that it was all on the record, you know, GPT 5 and some of the
- 0:22disappointments around that, which are very real.
- 0:25And also at some point said something to the effect of the
- 0:27quote in front of me basically like, yeah, I think we're
- 0:29probably there's some exuberance.
- 0:30So we're in a bubble right now where some stocks are are very
- 0:33overvalued because of AI. But then in the next sentence
- 0:36also was like but also AI is transformative.
- 0:39It's a classic power move. Elon does this too.
- 0:41It's like you're in the Super hypey thing.
- 0:43There's almost nothing to lose by being like, yeah, I don't
- 0:45know, People are really think I'm hot shit.
- 0:47Maybe too much, you know? Yeah, yeah, yeah.
- 0:50My, my, my lovers love me too much.
- 0:51My haters don't get it. But if there's something wrong,
- 0:54it's because of something other people did, right.
- 0:57I just want to disregard almost everything Sam says these days.
- 1:00I mean, it's just he's on. He says so much and it's so
- 1:04contradictory that it just doesn't even even though it
- 1:07probably did play some part in the sell off of AI, you know, AI
- 1:11driven stocks, the idea that oh, Sam Altman has declared a
- 1:14bubble, so we must be near some sort of a top.
- 1:16And then there was this MIT survey that came out and I think
- 1:19that probably had the most effect on a lot of people.
- 1:22Which headline number was that 95% of enterprise AI, generative
- 1:27AI projects fail just have 0 impact on a company's efficiency
- 1:31and bottom line. And that is it is a bad number.
- 1:35The problem, of course, is that a headline number versus what
- 1:38was actually in the survey is like pretty different.
- 1:41And this was not like a a survey that was put out by a fully
- 1:45disinterested organization like it's effectively a pro AI.
- 1:49Right, it's. Organization.
- 1:50It's always funny when somebody like runs away with one line
- 1:54that you said, but then like you're like, oh, but I'm sort of
- 1:58trying to make a pro AI point here.
- 2:01Yeah, yeah. Like it's almost like a rat fuck
- 2:04in that sense. Like if you're like one of these
- 2:05people who decides to buy the headline and then you actually
- 2:08like distribute the survey and then anyone actually reads it
- 2:10sees that, Like, the key points that it makes is that AI is not
- 2:14incapable of transforming a business.
- 2:17Just the implementation of AI and its inability to learn and
- 2:21adapt to people's workflow is what's causing it to be
- 2:24unsuccessful. My take away from actually
- 2:27reading through this thing quickly, quickly was that like
- 2:29the 5% of companies that have been successful with it are like
- 2:32going to win. Like it's really about like the
- 2:34fact that they implemented it better and were able to use this
- 2:37technology is why like it is actually a transformative tech.
- 2:40And like one of the co-authors of it is a product manager at
- 2:43Microsoft. I was just looking him up.
- 2:45So like these people that are writing this have a they have
- 2:48like a dog in this fight. And it's not like long term
- 2:51it's. Going to be a good flow of the
- 2:52argument is basically everybody sees the potential of Chachi BT
- 2:57and Claude. They try it out, but then they
- 2:59almost let and they just don't do enough to actually get their
- 3:02data integrated into whatever AI tools they're bringing to their
- 3:06companies. And so then it ends up not being
- 3:08more useful than just like consumer grade stuff and like
- 3:11here's what you actually need to do to be better than just like
- 3:15using Chachi BT. Right.
- 3:16Like that was the key point. Here's what you need to be doing
- 3:19to make this successful. And that's going to end up
- 3:21probably being the talking point from all of the sales people at
- 3:25the cloud software companies that are selling this stuff is
- 3:28like, oh, yeah, you read that report.
- 3:29They made these mistakes. You won't make this mistake
- 3:31because you're going to buy it and use it in this way.
- 3:33So it ends up the whole thing was an op, I've decided.
- 3:36I will say I texted a lot of investors this week just like,
- 3:40what's your read on the bubble like, you know, people who are
- 3:44like, you know, investing in these companies often.
- 3:46I'm interested in what, you know, yeah, people who are
- 3:49trying to make money like, you know, it's like if it's an early
- 3:53bubble, it's like, oh, I need to keep going.
- 3:54But you know, they're they're trying to time the bubble for
- 3:56financial reasons. And I did get some, you know,
- 3:58somebody who was like full bubble.
- 4:00Now trying to decide whether this is the start of 2021 or
- 4:03late 2021. Obviously, everything unwinds
- 4:06from the pandemic at the end of 2021.
- 4:08There's still upside to be seen at the beginning and AS.
- 4:12Is the person selling? They are.
- 4:15I am in tread lightly mode right now.
- 4:18Ha ha. The somebody else was like, I'm
- 4:21watching the Fed. You know, a lot of people are
- 4:23like, I don't know, what do you think?
- 4:26Yeah. I mean, so nobody's like, what
- 4:27an absurd question. You know, Andreessen Horowitz,
- 4:30who sort of dispositionally is like, why would you ever sort of
- 4:34position yourself as a bear? Our job is to fucking be
- 4:37professional. Bulls Just came out with a piece
- 4:39that was, you know, knocking down basically, I thought Tom,
- 4:43not trying to fight with your article about cursor.
- 4:45I think, I think the sophisticated critical.
- 4:48MBA so they just view me at the end of.
- 4:51It the sophisticated critical take, I think as you know, we're
- 4:54about to say tomorrow, it seems like foundation models revenue
- 4:57is still strong, but you know, application companies, like all
- 5:01the coding companies, you know, have pretty negative margins.
- 5:04They're built on top of money losing businesses and themselves
- 5:07lose money. How's that all going to work
- 5:09out? And Andreessen and Co, well, you
- 5:10know, I think it's, I assume it's Martine Casado and I think
- 5:14Sarah Wang sort of, you know, pushed back against your take.
- 5:17What do you think about that as somebody who is worried about
- 5:20app company margins, I. Think the margins are less the
- 5:23issue then like I think there's an interesting dynamic which I
- 5:27tried to get to in that story, which is the relationship
- 5:29between the model providers and the application companies.
- 5:32And it's clear with cursors grossly you can growth that you
- 5:35can build a huge user base on an interesting application, maybe
- 5:40has some brand name loyalty to it.
- 5:42But like especially with developers, like they will just
- 5:45switch to the thing that provides them the most bang for
- 5:47their buck. That's going to be the best
- 5:49quality product. And like the growth of Claude
- 5:51code over like 6 months, which I know wasn't impressive to Ed,
- 5:55but but is to me, you know, getting to like a $400 million
- 5:58ARR shows that like people will gladly switch to a new thing
- 6:01that seems like it'll work well. And Anthropic has, I mean,
- 6:04they're money losing too, but they're going to have better
- 6:06margins than Cursor will because they're paying wholesale access
- 6:10to the model versus the retail prices that that that cursor has
- 6:13to pay. So I'm not incredibly bullish on
- 6:16application layer companies in a world where model companies
- 6:19build the same software. If they decide they don't need
- 6:21to, that they really think they can make enough money through
- 6:24their API business or like a consumer, a pure consumer play,
- 6:27then maybe there's a world for application companies.
- 6:30But like, I think it's a very in this particular sector, it's a
- 6:34very narrow Moat. And I mean, I was talking to and
- 6:37I. Sort of agree with the
- 6:38Andreessen people. These companies are not idiots.
- 6:40They're gonna figure out like some of the stuff, like the very
- 6:43expensive stuff they're gonna offload to cheaper models.
- 6:46I agree with you that like the big business question is just
- 6:49like how sticky are these products?
- 6:50I think that yeah, it's like how sticky are they?
- 6:53I don't I and I don't, I don't have an answer, but I don't
- 6:55think the bubble is gonna swing either way.
- 6:57It's like, oh, they're not sticky.
- 6:58They'll because the people are moving to another AI product.
- 7:02You know, at the end of the day, coders want these coding tools.
- 7:06The margins will be figured out. They're going to consume some AI
- 7:09product, either foundation models or companies built on top
- 7:12of it. Either way, it's good for the AI
- 7:14hype train. Well, and I to defend the piece
- 7:17at least like I wasn't. Yeah, No, no, I don't.
- 7:20I don't just think too Andreessen because apparently
- 7:21they wrote it at me. Look, I think you know that that
- 7:24the cursor is not running out of money.
- 7:26Like the the story was very clear that, you know, yeah,
- 7:28negative gross margins are bad. So they have like a billion
- 7:31dollars or something? And they're losing, you know,
- 7:33apparently their burn is like in the single digit millions.
- 7:35So it's really an issue of like, you have plenty of time to
- 7:38figure these things out. And I mean, you mentioned Uber.
- 7:40Like I covered Netflix during its most cash burning days where
- 7:44I'll never not bring this up on the show.
- 7:45Like I worked for an editor who like, made a career for a time
- 7:49writing columns saying Netflix is a House of Cards.
- 7:52Right. Has he written me a culpa?
- 7:54No, we never do. We move on.
- 7:56We're like the we're like the Uber haters and the AI.
- 7:58Haters move on, you don't played it.
- 8:01I was the factual I provided. Where's your music pulper about
- 8:04Uber? Like, I was like the gun runner
- 8:06of the Uber skepticism era, you know what I mean?
- 8:09I was like giving all the haters, they're like ammunition.
- 8:12They needed to like shit on Uber's like money losing.
- 8:14But I was always like, I don't know.
- 8:16And then I always got annoyed that the people who took my
- 8:19stories and like, oh, they lose a lot of money did all these
- 8:21like insane math formulas based on them.
- 8:23And I'm like, there's no way. My numbers are like, so like
- 8:27there's such impressionistic, you know, like Uber was giving
- 8:30very like rough numbers to people when they were trying to
- 8:33back out their economics out of it.
- 8:34I was like, this is a lunacy. Yeah, well, the problem with the
- 8:37hater argument there is it it relies on the investors who see
- 8:40more than we ever get to really just being dumb, right?
- 8:44I mean, it's the idea that, like, they don't understand it,
- 8:46but we as the skeptic and the hater do.
- 8:47And believe me, I don't like giving investors credit.
- 8:50Like, it's not fun. It makes my job worse.
- 8:51But like, they do kind of do it for a living.
- 8:54Like that's the one thing they should be able to understand.
- 8:57Right. Is like the basic economic We
- 8:59should get haters on here too. I want to.
- 9:01Get some more haters, right? No, no, we like the Ed episode 1
- 9:03of the things the Youtubers. Somebody got mad at me for
- 9:05saying do you believe in capitalism?
- 9:07You know, like it was like some sort of shibboleth, but it was
- 9:10it's sort of a it was a sincere question.
- 9:12Like some of these people. I do want to understand like
- 9:15what premises do we share? Because like, to me, part of
- 9:18believing in capitalism is like investors are fairly rational.
- 9:22They're making different bets. They have their own, you know,
- 9:24it's just like I, I, it's useful to know how much people will
- 9:27think like the whole capitalist thing doesn't ultimately direct
- 9:30us towards likely outcomes. Like to me, markets are great
- 9:34truth mechanisms and like we play a part in that by, you
- 9:37know, point poking at things and interrogating them and then
- 9:41people change their ideas and then they eventually move their
- 9:43money. But when people have commentary
- 9:46that acts like investors won't listen to a good idea and, like,
- 9:49change their strategy if it's in their incentive, yeah, it feels
- 9:52like. Well, I mean like market, the
- 9:54truth mechanism is over time. I mean, like, the periods of
- 9:57irrationality in the midst of a bubble can give people the wrong
- 10:01impression of, like, the, you know, solidity of a business.
- 10:04Like, I think that's what Ed was arguing.
- 10:06It wasn't like capitalism is failing because these companies
- 10:09exist and keep raising more money.
- 10:11It's just like it hasn't reached the point yet where like, you
- 10:14know, whatever the Warren Buffett line about, like the
- 10:15shore, the shoreline going out and, you know, finding out who's
- 10:18naked. Right, right.
- 10:20And I mean, one thing we really believe in and I think smart
- 10:23business journalists are like there are people who can be
- 10:26behaving according to their incentives and driving us off a
- 10:29Cliff, right? I mean, this is sort of the like
- 10:31Tiger Globals and the soft banks.
- 10:33Like often I've made the point in the newsletter that it makes
- 10:36sense career wise to be the like biggest bull of the biggest
- 10:41mania, right? You're like, oh, at least they
- 10:43were the most important person, you know, like in that moment,
- 10:46like the Tesla longs and stuff, like what's her name?
- 10:49Kathy Wood? Like, even if you have a
- 10:51terrible performance, you sort of like win fans just for being
- 10:54like really extremely bullish. And so even if you like believe
- 10:59in capitalism like I do, certainly I think they're all
- 11:01these actors who are like, you know, almost like taking
- 11:04advantage of, you know, human psychology rather than, you
- 11:07know, trying to make the most money they can.
- 11:09They're just, they're like, oh, I'm going to be able to, if I'm
- 11:12the best bull in the world, I'll be able to raise the most money
- 11:14again in the next bull run in my game is just to like run as hard
- 11:17at the bull market as possible every time.
- 11:20And you know, like that's a, that's a game people play with
- 11:23with great success. Speaking of taking advantage of
- 11:25human psychology, A Chamath is back out there again.
- 11:30Welcome back. Scam in the arena?
- 11:32Scamming the arena's back? No crying at the casino Chamath.
- 11:35Right. I know what a what a he's so
- 11:37good at it. He's so good at it.
- 11:38He's like, like, he's basically calling himself like the casino.
- 11:42And it's like, sure. The casino the the the random
- 11:45outcome generator I understand. Don't.
- 11:47He's literally like, don't invest in me if you're not
- 11:49willing to lose all your money. And if it's like trivial to you,
- 11:52like, yeah, who invests like that even if you're rich?
- 11:54No rich person wants to lose their money either.
- 11:56Like everybody is investing to get a return.
- 11:59Like, I know it's just like a bullshit thing you say, but.
- 12:01This is the unique twist of the Trump era, right?
- 12:03Is that like he has discovered there is a whole cohort of
- 12:07Americans who like being scammed, who like.
- 12:10Well, that's true. I mean, sports betting is the
- 12:12best thing in the world. Like yeah, exactly Like, yeah.
- 12:14I lost a lot of money on sports betting over the weekend.
- 12:17Yeah, Yeah. God, yeah.
- 12:19That was just fun. It was.
- 12:20It was. Just for the fun of it.
- 12:22Was just for the fun of it. Yeah, there.
- 12:23I just went down to the sports book and I put $50 on a on a
- 12:27horse called Duck Stuck go and at.
- 12:28Least it was in person. You got some, you got an
- 12:31experience out of it. My my issue with the sports
- 12:33betting apps is like, what are you even getting?
- 12:35You know, like. I think it's really mistakes of
- 12:37watching a game. I mean, if you're not watching
- 12:39the game, you're putting money on it, then that's a real,
- 12:41that's real sick, right? I did invest also.
- 12:43I put a bet on Australian rules football because I thought the
- 12:46spread was so ridiculous, but turns out it was not that team.
- 12:49That team lost by like 150 points, which I don't know what
- 12:52that actually means. Anyway, Chamath is yeah, I think
- 12:54he absolutely represents the peak of Trump era.
- 12:57We like to be scammed mentality where he is actively calling
- 13:01what he's doing a casino that there will be crying and you
- 13:04will not make any money. I will make money from it.
- 13:06He made money last time. I mean you like wrote I thought
- 13:09a great piece about like his scam the last time through.
- 13:12Like what are you expecting this time around 2:00?
- 13:14Pieces on Schmoth. I mean, one thing that makes me
- 13:17so sad about the world we live in is just how much your
- 13:20Internet brand and your real person brand can be so far
- 13:24apart. Like real people in Silicon
- 13:26Valley think he's like a huckster.
- 13:28Like people who work with him, it ends poorly for them.
- 13:30Many of his professional relationships have gone poorly.
- 13:34He didn't make people money on the SPAC's.
- 13:36He made money through fees. Like he's looking out for
- 13:39himself. And to some degree, you still
- 13:42want somebody who's like looking out, who cares about caring
- 13:45about your reputation. He's good because it means that
- 13:48the part people you partner with and work for and invest their
- 13:52money for you, you care about your long term reputation.
- 13:55You want to be good by them. And this is someone who hasn't
- 13:58minded his long term reputation. It hasn't, you know.
- 14:01It's not a reputation as an investor, but if you look at the
- 14:03way he positions himself, he wears his like, I don't know the
- 14:07fucking brand name suits. He loves to be rich, yeah.
- 14:10Yeah, And so, like, again, and not to make everything just
- 14:13about Trump, but this is all part of the same thing.
- 14:15People in the real estate world knew that he was an absolute dud
- 14:18as an investor. Like all of his projects would
- 14:21go bankrupt. He was a brand name to to, you
- 14:23know, add a modicum of flash on top of an otherwise shitty
- 14:26project. Here you're like, you're like,
- 14:29bring back the Wasps. Where were the rich people who
- 14:32cared about I? Mean what?
- 14:34Old money. Just the people who.
- 14:35Cared in their wealth. No bless oblige.
- 14:37You know, we need, we need like people who are not.
- 14:40We just need the Carnegie. I am watching Gilded Age right
- 14:42now. So I have a very strong
- 14:44attachment to to that era of, of, of wealth.
- 14:47But yeah, those guys look tremendous compared to the
- 14:49Chamats. I mean, they've done nothing or
- 14:51he's done nothing, but he's built an an image that I'm sure
- 14:55he will have no problem getting people signing up for his specs.
- 14:58Well, it's that sort of fatalism that makes it possible, right?
- 15:00It's just sort of like people are like, oh, he'll raise the
- 15:02money. We should do it and.
- 15:03Yeah. And I guess maybe the last point
- 15:05to. Talk about I'm not cheering for
- 15:06him. I don't know what to say.
- 15:07I would be very sad if good people lose money on what are
- 15:11certainly going to be failed IP, OS or whatever you call them.
- 15:14Specs. I do think to be fair, you know,
- 15:17we're business journalists. Like I do think he changed his
- 15:20structure this time. I do think he's slightly more
- 15:22aligned now with the actual outcome of the specs, though.
- 15:26Knowing Chamoth, he's found some other way to make money as well.
- 15:29But. Burdens.
- 15:31I don't see why, but he's not scamming people, sorry.
- 15:34This. Well, that you believe in
- 15:36capitalism, don't you? There's infinite numbers of
- 15:38shots on in a free model. Sure, yeah.
- 15:40But I mean, this is always, this is one of these, I don't know,
- 15:43to me psychological weaknesses the the elites.
- 15:46Once you're in the club, you're like, oh, you're in the club,
- 15:48you should get to do it again. It's like there there are tons
- 15:50of human beings like try it with somebody who hasn't lost people
- 15:53a lot of money. Like to me this sort of desire
- 15:56to like, keep giving money back to the type of people, to people
- 15:58who just like, gamble the money rather than saying maybe we
- 16:01should try it with somebody new, that it means like a human
- 16:04psychological weakness of just like loving fame rather than
- 16:08success. It's never gone away.
- 16:10I mean like Michael Milken in the 80's, the junk Bong King
- 16:13still like runs a large financial conference in Los
- 16:16Angeles, the Milken conference. Nothing, nothing tends to really
- 16:19happen to these people for the long term.
- 16:20I mean, some of them do get like banned from trading, but that's
- 16:23not even close to what's going to.
- 16:24Happen with, not anymore. Yeah, exactly.
- 16:26That's it. Like I'm not accusing Chamath of
- 16:28anything of that level. He just was an opportunist and
- 16:30fucked over a lot of people who didn't.
- 16:34We're looking at fights. So Chamoth come on the show.
- 16:36We're I don't think he is. That would be a big get for lots
- 16:40of reasons. Is probably the person who's
- 16:41written the most negatively about him.
- 16:43I'm not betting on that one. But if you're if you're a hater
- 16:46positively negatively, we're we're ready to argue with you.
- 16:48We like this arguing format. So we we want to have arguments.
- 16:52Not quite an argument. I like him too much.
- 16:54But after this we're going to have Guillermo the Guillermo
- 16:57Roush over cell. He's coming on the show.
- 17:00I think he had, I know a really good prediction about where
- 17:04things are going. We talk about agents.
- 17:07He obviously runs a no code company and I continue not to.
- 17:11I have not built my no code startup empire.
- 17:14If it's so easy to no code like where, where are the apps that
- 17:17I'm? Using maybe you should start
- 17:17getting into fights with lovable.
- 17:19I know dude, I don't know if people know what you're talking
- 17:21about. We don't yeah, that that can be
- 17:23deep lore for newcomer fans. You can look up Eric's tweets
- 17:26trying to get his. Money waste your time for I I
- 17:28lost it on Sunday just because I I got charged twice from lovable
- 17:33after trying to cancel. They, they have extreme dark
- 17:36patterns that they say they're going to try to reform.
- 17:38Literally you click a big red button that says cancel and then
- 17:42you still are not cancelled from lovable.
- 17:44They're like, oh, that was cancelling within lovable.
- 17:46Then we had to send you to stripe and then you have to
- 17:48obviously run through the stripe.
- 17:49But anyway, they, they swear they're changing.
- 17:51They're going to fix it. Maybe cancel in Swedish means
- 17:54something else? Maybe it means like I'm still
- 17:56down. I like the.
- 17:58Company I would have come I it was just like I'm not using it
- 18:00this second. I would have come back now.
- 18:02I don't know. I'm pretty angry with them, but
- 18:04hopefully they're. Reformed we're big supporters of
- 18:06Europe tech over here on newcomers so we we long term
- 18:09long term bulls on lovable and every European tech company all
- 18:13right we should probably cut on over to your interview with
- 18:15Guillermo all right I'm. Here with the founder and CEO,
- 18:20Versal, Guillermo Roush. Welcome to the Newcomer Podcast.
- 18:24Thanks for having me. Last time we were hanging out,
- 18:27I, I, I feel like you convinced me that Purcell would be every
- 18:30company in the world. So we'll see.
- 18:33We'll see what you've brainwashed me into by the end
- 18:36of this conversation. Where?
- 18:38Where were we? I I think it was huminex if I'm.
- 18:42I'm remembering, I think we. Had like lunch off that and just
- 18:44like the the great founder visionary that I was like, oh,
- 18:48what, what won't this company do?
- 18:51You're you're obviously touching a lot in AI.
- 18:54You know, like, you've got VO now with sort of your own vibe
- 18:57coding product, but you're also sort of powering a lot of like
- 19:00the vibe coding that everyone's talking about in sort of some of
- 19:03the infrastructure to make it possible for people to sort of
- 19:09spin up sort of websites and applications quickly using
- 19:12artificial intelligence. But also, you're just like a fun
- 19:15founder who's willing to sort of like, say what's going on and,
- 19:18you know, not, not just let the venture capitalists dominate the
- 19:23chattering among the Silicon Valley classes.
- 19:25And terminally online so you can ask me about any of this
- 19:28happened on X5 minutes ago and I'll know.
- 19:31So I, I just want to start off with, yeah, like the existential
- 19:34question about like the mood. I mean, I've been honestly been
- 19:37texting everybody today about like sort of animal spirits or
- 19:41like how's the, how's the market?
- 19:43Like, you know, what's your read on sort of the AI mania?
- 19:46Like, do you feel like there's still a lot of room to run here
- 19:50or are you getting signs that there's like some some paranoia
- 19:54about the euphoria? You mentioned Versailles is in a
- 19:58very specific and special position that we see every,
- 20:03almost everything that's being built on.
- 20:04AI has to have some kind of user interface to the world.
- 20:08And so traditionally, we've been supporting founders in building
- 20:12this AI portals, right? AI applications now increasingly
- 20:16agents. And, and one of the things that
- 20:18I've been noticing is that there's basically 3 chapters of
- 20:21AI so far. Number one was assistance #2 and
- 20:26it's the arc that I think we're currently in is agents.
- 20:29I believe that there's a third phase.
- 20:31I think what's really interesting being so close to
- 20:32developers is that I've always kind of called out us
- 20:36developers. We're very self-serving.
- 20:39So if we get AI, we're going to apply to our job first.
- 20:42We're going to make our own life easy.
- 20:45And So what I'm seeing with developers is that the next
- 20:47generation is going to be teams of agents or multi agent
- 20:50architecture. We've gone from assistant to
- 20:52agent to multiple agents. The alpha developers today, what
- 20:55they're doing is they're, they're instead of just limiting
- 20:57themselves to like 1 agentic session, they're they're
- 21:00spinning up 20 in parallel or 10 or 5 or whatever.
- 21:05And that gives you a glimpse of what I think is going to be the
- 21:07next phase of AI. You can think of the current
- 21:09phase of AI as you have a virtual Co worker.
- 21:13You can think of the next phase of AI as they have a virtual
- 21:16team of Co workers and I'm the manager.
- 21:19Is the challenge with just having the virtual Co worker is
- 21:21in some ways the human gets the lame part of the job?
- 21:24Or it's like I have to sit there and be like, sure, yeah, I keep
- 21:27keep doing the hard work. Like, yeah, I keep just
- 21:29hallucinating. Yeah.
- 21:30Like, and it's tedious and, and perhaps you don't get that 10X
- 21:33effect. You know, I do see a lot of
- 21:35debate, healthy debate sometimes of like, is AIA 2X or is it a
- 21:4010X for my job? I saw an article that a
- 21:43programmer put out the other day saying like, look, in certain
- 21:45areas of my job, it definitely feels like a 10X because the
- 21:48example that this gentleman was giving was like, I don't write
- 21:51compilers every day. And so like the fact that I was
- 21:54just able to sell the AI, can you write me this like compiler
- 21:57transformation is like, whoa, like I feel like a superhuman.
- 22:00But then he was saying, like in other areas of my job, like, I
- 22:03don't know, sometimes it like kind of gets in the way and
- 22:05whatever. To to sort of really, I guess go
- 22:08directly at the current moment. I mean, yeah, what's your
- 22:11reaction been to like ChatGPT 5, I guess like to GPT 5 because in
- 22:17some ways, I guess I think if that were amazing, the lead here
- 22:20would be like, oh, the models are going to keep coming up with
- 22:23new ideas and that's where we're going to make progress.
- 22:25And just by the fact that you're saying, oh, we're going to find
- 22:28ways to use the existing models and agents to like get more out
- 22:31of them by having them collaborate to me as sort of a
- 22:33sign that we're not having the like, Oh my God, Chachi PT or
- 22:38GPT 5 is like blown all our minds.
- 22:40Do you agree with that or what's your read on it?
- 22:42I agree with you that analyzing the GPT phenomenon is really
- 22:46important to the industry because you're right.
- 22:48Like it did, it came out and it didn't come with a short proof
- 22:52of the Riemann Hypothesis. Exactly.
- 22:55Where's our exactly in the expectation or whatever?
- 22:59Yeah, right. But something that was really
- 23:02interesting was that Open AI is taking the coding problems
- 23:06really seriously prior to its release.
- 23:09They came to us and they're like, can you please run this on
- 23:11V-0? Can you please run your evals
- 23:12with it? They were asking us like what
- 23:14are the vibes from your team? What are the vibes from your
- 23:18tests? Is it good at design?
- 23:20So there was a lot of attention being paid to the topic of
- 23:24coding. And the simplistic view of that
- 23:28is that the market is becoming hyper competitive on the
- 23:31enterprise side and that Anthropic is making a ton of
- 23:33progress with coding. There's amazing Chinese models
- 23:36coming out like Quen, etcetera. They're really good at coding.
- 23:40And so the simplistic view is like, oh, coding is a hot
- 23:42market. So that's what they're going
- 23:43into it. My view is actually that because
- 23:46agents are the future, not assistants and not one off one
- 23:51shot things, the future to superior and and and higher
- 23:55intelligence and more productivity will be agents
- 23:58writing code to solve problems. So the thing to pay attention to
- 24:01is not a consumer observation. In fact it almost backfired for
- 24:05consumers right? Because we were like, can you
- 24:08please bring back my boyfriend or girlfriend or?
- 24:10Oh, right. We don't want it to be smart.
- 24:11We just want it to be. Yeah, like that was the main
- 24:14thing, right, you know? Art of Maine had almost
- 24:17anticipated that this was going to happen because they did the
- 24:19following thought exeriment. If we could revive Albert
- 24:23Einstein and we all get a digital copy, how many people
- 24:27would be jazzed? I know a lot of people here in
- 24:29Silicon Valley. You and I would be like, holy
- 24:31crap, right? Einstein in a box.
- 24:33Sort of like you're a crazy guy you don't want to talk about
- 24:36like, reality television or whatever.
- 24:37Like when I wouldn't want to talk to you all the time.
- 24:39I see the point, right? Yeah.
- 24:41The average person would be getting too much horsepower.
- 24:44They're like Kate Albert. Like, what should we watch on
- 24:45Netflix tonight? And he's like, I don't know,
- 24:48like I don't watch Netflix. Let's talk about internal
- 24:50relativity. And so I think GBT 5 had a
- 24:53little bit of that of like, it's becoming intelligent in a way
- 24:56that's going to enable this like emergence of intelligence that
- 25:02you can't easily probe for. The future of intelligence will
- 25:05be and, and we're seeing this with like test time models, like
- 25:08going in a loop, spending a lot of energy writing code, testing
- 25:13it, exploiting multiple branches of the search algorithm.
- 25:17And these are things that a consumer cannot possibly
- 25:19ascertain. Like the most that a consumer
- 25:21can do is, you know, I have friends that keep questions in
- 25:25their head that they know like a is are not very good at.
- 25:29And so like they do their own little benchmark of like, oh,
- 25:31let's see if GBT 5 gets this right now, testing the
- 25:35intelligence or the IQ of AI in itself is becoming really,
- 25:39really difficult. That's why I open AI was so
- 25:40interested in like, what is what's the B0 point of view?
- 25:44You're creating a coding agent that you know is a specializing
- 25:47in in vibe coding, etcetera. And so that means you have a lot
- 25:51of tests. That means you have a lot of
- 25:53benchmarks like, you know, what does that mean for you?
- 25:57I believe every company should be concerned with creating this
- 26:00benchmarks of their own institutional knowledge and
- 26:03intelligence. Well, some of it, you know, it's
- 26:05like for me, it's like I put it in like, oh, is it good?
- 26:06How good is it now at writing a story like me?
- 26:09How, how you know good? Is it keeping track of yes,
- 26:12proofread problem. You know, you, you end up having
- 26:14your specific challenges. So when you're in an industry,
- 26:17it's not that hard to figure out like, oh, is it doing a better
- 26:20job for me or not? You could also make the case
- 26:22that you should, you could be working on a writing agent that
- 26:26it's more than just one forward pass of like.
- 26:29And this is another thing that kind of backfired, but it's also
- 26:32very interesting is that they try to ship in a model router,
- 26:36right? Like the default thing was auto.
- 26:39And then it's going to choose if it's going to think hard or if
- 26:43it's going to give you an immediate response.
- 26:46Another thing we've learned about AI that's fascinating to
- 26:48me is that there is this metaphor of like thinking fast
- 26:52and slow, for which there is a book and like people have been
- 26:56talking about for years, like prior to like AI existing.
- 26:59And it's, it's proven true. There is 2 modes of thinking.
- 27:02There is fast, fast twitch muscle fibers or like you just
- 27:06want like, hey, like, can you provide this really quick and
- 27:09find me typos? And then there is a like, no,
- 27:10no, no, like, let's analyze. Like, are you repeating words
- 27:14too much? And that requires like a slow
- 27:17thinking. And so open AI try to, because
- 27:20they're in the pursuit of creating a good consumer
- 27:22experience. They were like, OK, we're going
- 27:23to try to route automatically. That appears to me to be an
- 27:28extremely hard problem, if not impossible.
- 27:30It's hard to know from the query if it's a hard problem or a slow
- 27:34problem. We have a related problem in
- 27:37computer science called the halting problem.
- 27:39Like by just reading some code, we don't know for how long it's
- 27:42going to run. It could run until the death of
- 27:46the universe, or it could run really quickly, but we have to
- 27:48run it to actually know. And so there's almost like an
- 27:52uncertainty of like, should I spend a lot of energy thinking
- 27:55here or should I respond really fast?
- 27:58In some ways, this is the argument for the role of
- 28:00startups in AI and in sort of narrowly focused companies.
- 28:05It's like, really, you're saying the model companies should
- 28:07almost be an API layer? They're not going to be great at
- 28:10deciding in every sector whether it should be a faster or slow
- 28:13thinker and or what. Content or how many agents you
- 28:17use, etcetera. And so where I was going to go
- 28:18with your case is that it'd be really cool to be crafting the
- 28:23ideal writing agent. I, I know most people listening
- 28:26to this will have an intuitive sense of the answer to this
- 28:29question, but like, I'm curious to hear you define it.
- 28:33What, what is an agent to you? Like, does it need to take
- 28:37action? Like, I mean, deep research
- 28:38itself is an interesting case where it's like, oh, I mean, a
- 28:41consultant is like an, an, you know, taking actions in a
- 28:45certain way and deep research behaves like a consultant.
- 28:48On the other hand, and people have probably heard me say this
- 28:51before on this podcast, Like to me, the gauntlet for an agent is
- 28:54like, can you go spend my money? Like, will I let you really make
- 28:57sort of permanent sort of decisions that that have real
- 29:01consequences in the world? I don't know.
- 29:03How do you think about what what sort of the minimum requirement
- 29:07for? What makes something the
- 29:08minimum? Requirement is that it produces
- 29:12an artifact, an output that is generally longer than an answer.
- 29:18So it right, like if, if I, I make that decision between like
- 29:23an assistant and an agent. So if I'm asking like, you know,
- 29:26what's the province main the largest province in Argentina
- 29:32versus build me an application. Funny enough, for us, that was
- 29:36like one of the biggest jumps in how people have perceived the
- 29:40intelligence of B0B0 started being what do we call one shot?
- 29:44You would tell it, give me a user interface for a document
- 29:49collaboration system. And the feedback that we got, it
- 29:52was like some people like literally have because this was
- 29:54so early, like it was basically maybe a few months after Chad
- 29:59GBT launched, some people would tell us, all right, there's not
- 30:04going to be any jobs anymore. I just saw God, like, holy crap,
- 30:08what do you guys do? And some people tell us this is
- 30:11the worst piece of crap ever because on in the one shot world
- 30:16quality so divergent. It's almost like playing the
- 30:19lottery. In fact, some naysayers of AI
- 30:22for coding are saying that like what drives the revenue is not
- 30:26the quality is the lottery effect, which I completely
- 30:30dismiss. But I wanted to tell you this is
- 30:31funny. It's a dangerous idea because
- 30:35yeah, it's a grabby idea. Interesting.
- 30:37The next big jump for us was when we made V-0 really work
- 30:42like an agent. And So what it does is that it
- 30:45makes that one shot, but then he looks at it, Visio asks itself,
- 30:51does it work? Does it compile?
- 30:54The most recent version of her agent even physically looks at
- 30:57it like using computer vision. So as an example, say I want to
- 31:01make a website for my daughter's birthday.
- 31:04Please theme it like, I don't know, Disney Princess.
- 31:10It'll then look at what it cooked.
- 31:11It's like, does it look like what I was asked?
- 31:14And so an agent is taking a multi step approach, much like a
- 31:19human would in order to produce a higher quality output.
- 31:25One of the key ingredients there is this idea of reflection, like
- 31:28you're the agent is asking itself questions, is looking at
- 31:31artifacts. It is it's using tools.
- 31:34One tool that we have is it can look at a screenshot, it can
- 31:38look at designs for inspiration. So sometimes the model doesn't
- 31:42know what you mean by like build me a podcasting app like
- 31:45Riverside and so he needs to go and research.
- 31:48Do you like the term vibe code? By the way?
- 31:50Do you embrace that term? I personally do it's funny
- 31:55enough because like our products becoming so successful, like the
- 31:57Fortune 500, some people are like, do you really want to say
- 32:01vibe coding? I think vibe coding is it's it
- 32:06evokes this idea of you're not actually coding, which is really
- 32:11important. The term came out because Andrew
- 32:15Karpati and a bunch of our people were realizing AI was
- 32:19progressing at a clip where your own micromanagement of the agent
- 32:24was producing diminishing returns.
- 32:26AI Once Upon a time, like in AI times like 6 months ago.
- 32:30Once Upon a time it was so bad that you literally had to look
- 32:34at every token it output. As the models get better and the
- 32:40world becomes more agentic and self healing, you realize that
- 32:44you're looking at it. It's just wasting your own
- 32:46energy and your time. You can start trusting and even
- 32:51seeing because we can render that output.
- 32:54We can run the application. You can see it come to life.
- 32:58And so there has to be a term for that.
- 33:01The winning term right now is by the coding, but it's basically,
- 33:05sometimes I've called it instead of being coded first like a
- 33:08product, like cursor, it's code last.
- 33:12You might occasionally maybe at the end of the process you look
- 33:14at the code, but the code is not the important thing here.
- 33:17And I believe that most agents are going to going back to like
- 33:21why GBT 5 is obsessed about coding, even though people are
- 33:24obsessed about four O being nice to them.
- 33:27I believe that agents will have to write code to get to that
- 33:31next frontier of value for society and for intelligence.
- 33:35And I believe that a lot of people will not even realize
- 33:37that this agents are writing code.
- 33:40What's happening to your employee salaries?
- 33:43I mean, you know, the the AI engineers, obviously we've seen
- 33:46all these stories of meta poaching everyone from the
- 33:50research labs, like how's this playing out for sort of the
- 33:53typical software engineer in Silicon Valley or what?
- 33:57Yeah. What are you seeing in terms of
- 33:59trends and comp? Are the the $100 million
- 34:03salaries as widespread as chattered about?
- 34:06We are in a little bit of a microcosm here, a metaphor that
- 34:10someone gave me their days like look like it seems like if you
- 34:13extrapolate what you're saying. And because I know that V-0
- 34:17hires developers to build V-0, we're transitioning into a world
- 34:22where the job of an engineer is to build the factories of code
- 34:28rather than being the line worker.
- 34:32And so the engineers that are capable to building the
- 34:34factories of code, the Bible coding platforms, the agents,
- 34:39the AI interfaces, the assistants, those people are
- 34:43highly empowered today. And so it becomes that highly
- 34:47competitive space. It's like, you know, how many
- 34:50soccer players like Lionel Messi are there that have that
- 34:54specific set of skills? And then like, you know, there's
- 34:57lots of soccer players, but a handful that make a ton of
- 35:01money. And So what I would recommend to
- 35:05people is that they position themselves in that world, right?
- 35:08Like become the person that understands how to bring those
- 35:12two worlds that I talked about earlier.
- 35:14The theoretical AI potential and it's applied AI presentation
- 35:20layer. That's kind of the, our thesis
- 35:22of the Versailles AI cloud is that as this model layer keeps
- 35:27getting smarter, better, more accessible, you should be
- 35:32focused on applying that AI to specific domain, a specific
- 35:36business, a specific vertical, a specific kind of agent.
- 35:40But you're going to need a cloud that has all of the like right
- 35:43patterns, tools, defaults, etcetera, for building those.
- 35:49And that's, that's our thesis with, you know, we're not going
- 35:51to build a cloud that is 1 to 1 exactly like AWS.
- 35:54There's no advantage there, but we want to build a cloud that is
- 35:58perfect for that next generation of companies and developers.
- 36:03I'm trying to understand. I mean, this is we're in an
- 36:05insane situation where you're right at the high end.
- 36:08If you're building the AI factories, you could command
- 36:11insane salaries. On the other hand, if you're
- 36:15sort of the sort of, I don't know, worker, more like line
- 36:19coder, software engineer, I guess you're worried like, oh,
- 36:22you see V-0 and you're like, oh, that's coming for my job.
- 36:25So my question I think is coming at you from 2 ends, which is
- 36:28like at once, like, oh man, we see these insane salaries.
- 36:31On the other hand, I feel like they're the sort of, I don't
- 36:34know, Wall Street Journal story saying don't, don't study
- 36:37computer science because they're not a new job.
- 36:39So I guess take one at a time because I think they're getting
- 36:42mixed. They're getting mixed together.
- 36:44One talk about each end of of the equation here.
- 36:47For the other end, yeah, I do have a complete different take
- 36:51on universities and colleges and that is a kind of worms.
- 36:55I think we should address it. OK.
- 36:57On the line coder, I also think that the demand for code from
- 37:01those people has actually never relented.
- 37:05Meaning let's say that you were able to produce like 1000 lines
- 37:09of code per hour per day, whatever your boss probably
- 37:15wanted you to produce even more like there's more features or
- 37:18bugs, there's more innovations, more products, etcetera.
- 37:21And so if you use AI, you're that's why I was saying that
- 37:25it's so exciting to be seeing people go into this multi agent
- 37:29ways of working. And so the person that uses AI
- 37:34is going to run laps around the person that doesn't.
- 37:37But do you? Think they have more like
- 37:39frontline engineers in a decade than we do today?
- 37:43It's very hard to estimate in, in like in total numbers, right?
- 37:48But I do think we'll definitely have more software, right?
- 37:51Like, right. Yeah.
- 37:52The demands, we all agree on that.
- 37:55The, the definition of an engineer will likely expand in
- 37:58such a way that I can, I can tell you definitively, yes,
- 38:01right. Like, because we're putting
- 38:04software generations in the in, in the hands, in the hands of
- 38:07everybody as a whole, there will be, you know, hundreds of
- 38:12millions, billions of people creating software, be a specific
- 38:16engineer as we understand them today.
- 38:19I can't even, you know, I, I had a tweet a couple days ago that
- 38:23Elon replied to say I, I was saying, like, I can't even
- 38:28explain just how different software engineering will look
- 38:30in five years because the perception that people have
- 38:35today is of a code editor. Like that's it.
- 38:39Like if you watch the hacker movie, what is the idea of a
- 38:42programmer or they're looking at a code editor?
- 38:45But now I'm looking at my fellow factory builders and I don't see
- 38:53them looking at code editor anymore.
- 38:55I'm increasingly seeing them talk to all these agents.
- 38:59And so that to me already like if I was able to transport
- 39:02myself to five years from now, I don't think software engineering
- 39:06looks anything like what do we do today?
- 39:08And so it's very hard to say like like are we calling those
- 39:11people that are managing the fleet of agents engineers?
- 39:15You know, most likely I would say it's, it's more on the
- 39:18customerization or democratization of software
- 39:20creation. To a direct point you were
- 39:23touching on earlier, do you think people should study
- 39:26computer science right now? I think people really need to
- 39:31understand how AI systems work, period.
- 39:36They need a deep understanding of the technology stack, the
- 39:39limitations of the technology, what makes them safe, what makes
- 39:43them not safe, what makes them make mistakes, whether they're
- 39:47good, whether they're are they not good.
- 39:50I do have a little bit of apprehension about do the
- 39:54universities today themselves know?
- 39:57And I'll say most likely not, right?
- 40:00Like, think about how fast the industry is moving.
- 40:03Right that. You're like come work in
- 40:05Versailles and figure it out here and skip our user.
- 40:08Road map is moving week by week, right?
- 40:12It's very hard to say in Q32026, Visor is going to look this way.
- 40:18We're going to need to find a way to get the universities up
- 40:22to speed on AI, on vibe coding or start to blur the lines
- 40:28between university and industry. We've been running very
- 40:33successful internship programs at our cell where I personally
- 40:36spend a lot of time with the interns.
- 40:37Like I'm picking their brain, they're picking mine.
- 40:40Like this intern is like, you know, hanging out with CEOs and
- 40:44CT OS because like, I do think that you are getting up to speed
- 40:48really fast if they're in the right room.
- 40:50I do think there's a little bit of a get rid of your ego.
- 40:53Forget about everything you are good at.
- 40:55It starts from first principles. If you, if you thought you were
- 40:59really good at TypeScript, if you thought you were really good
- 41:02at C, well, yes, but also someone that never knew how that
- 41:09technology even worked will probably be generating lots of
- 41:13it just like you very soon. They might be even be learning
- 41:18faster than you because you're able to try more things.
- 41:21I'm very concerned in the sense of like people say, the junior
- 41:24developer job. Instead, I would argue maybe
- 41:27there's nothing that's ever been more important than using your
- 41:31company with clinical junior developers that are more AI
- 41:35native and can show you new approaches truly to learning, to
- 41:41iterating, to figuring out how to solve a problem when you
- 41:44don't know all of the science behind it.
- 41:46When you ask about computer science, computer science is a
- 41:48very specific version of like how they teach you data
- 41:51structures and algorithms, etcetera.
- 41:54If they're leaving out this like pragmatic AI side of
- 41:58engineering, I really liked what Elon said about like we're
- 42:01nuking the term researchers at Grok, at XAI, we're all
- 42:05engineers. I do think it's a very healthy
- 42:07mindset to have. We're all engineers and we're
- 42:10all figuring it out and we're all using AI to figure out as we
- 42:13go. Taste has become, you know, a
- 42:16keyword in how people think about the role of humans going
- 42:21forward in AI. And I, I want to use that as a
- 42:25lens to sort of untangle some of the coding stuff we're talking
- 42:29about. Like, you know, obviously I
- 42:31spend a lot of time thinking about writing with writing like
- 42:33you can, everybody can have taste about stories because we
- 42:36can all like read and assess them.
- 42:39And obviously there are sort of the practitioners who sort of
- 42:42know what goes into making it. But then there have always been
- 42:45sort of critics of, you know, novels and writing that sort of
- 42:49you can imagine with AI, the sort of line between the writer
- 42:54and sort of this sophisticated critic of writing sort of is
- 42:58collapsing because you can sort of use AI to write over time.
- 43:03I guess with code. My my question is like, there's
- 43:07sort of the, there's the product, there's like what what
- 43:10is the thing outputted that you could have taste on?
- 43:13And that's something that seems like it's being democratized,
- 43:17like potentially writing. But won't this sort of code base
- 43:22itself still matter? Or like, will we need engineers
- 43:26that have like the taste for like, what is well written code
- 43:30and like how systems are set up? Or are we just going to like,
- 43:33yield that piece of it to agents?
- 43:36Like, is the taste only about whether the product sort of
- 43:39itself looks good, works well, or is it about like how the
- 43:44actual code base is set up? When I first came to Silicon
- 43:47Valley, one of the things that was and vogue was there were a
- 43:52handful of people in the Valley that could create deep tech
- 43:57software systems and also design beautiful products and I could
- 44:01count them with like one hand. There were a handful of people
- 44:06that I knew and I aspire to become like them.
- 44:09I've cared always a lot about the design and the engineering.
- 44:13To me, that's what was so special about the Steve Jobs and
- 44:16Apple mythology is that they literally had a slide in one of
- 44:20their presentations of like, Apple at its best is the
- 44:23intersection between computer science and liberal arts.
- 44:27And they had a sign of like the intersection of two streets, one
- 44:29called computer science, one called liberal arts.
- 44:33I think the future will be more of that.
- 44:36I would argue that the liberal arts is going to be overweight
- 44:40because the computer science part, well, the agents will
- 44:43figure out all the data structures, algorithms,
- 44:45optimizations, blah blah. And so your ability to infuse
- 44:49creativity, taste, culture, design will probably matter even
- 44:55more in the future. You know, I'm happy to hear the
- 44:58argument for like, the liberal arts major, but I still think,
- 45:02you know, tech companies still need to build things today.
- 45:05And it's not just about like clever billboards and clever
- 45:08marketing. Yes, but there is a discovery
- 45:11element to technology like there's just so much technology
- 45:16that is possible and yet we don't know it's available to us
- 45:20and so when people. Say this stuff, I'm like, oh, I
- 45:22have a distribution platform. This is what everybody seems to
- 45:24want. I just need to do what you guys
- 45:26are saying is the easy part. Oh, just cobble on some tech
- 45:29thing to sell people. It's like, oh, I have people's
- 45:31attention. It's like clearly I'm deficient
- 45:34in figuring out the, you know, tech thing to just like cobble
- 45:37on. But but to me, from where I sit,
- 45:39it seems harder than some of the vibe coding it takes.
- 45:43It's like, oh, you can't just spin up.
- 45:44You can't be like, oh, I have attention, let me just spin up.
- 45:48There is a software that I pointed there that you're
- 45:50pointing out, which is that even if you figure out a huge
- 45:52platform, you need to create something that's great for that
- 45:55platform. And so if software is getting
- 45:58democratized, like maybe it's when I mean, when I mean that
- 46:02everybody can cook. A lot of it is like solving your
- 46:05own problems and and creating tools for you and your
- 46:08colleagues and things like to that matter, creating software
- 46:12at scale, like if you get the biggest like Times Square sign
- 46:18and you're going to get every eyeball on the planet and you
- 46:20want to give them something of high value.
- 46:22It's so really hard. Yeah, you can solve your own
- 46:25problems. You know what you need.
- 46:26But to solve some sort of general problem for a bunch of
- 46:29humans is is a hard product. Building that is that's very
- 46:32hard. My argument was there's
- 46:35something really interesting about the Steve Jobs story
- 46:37because he wasn't an engineer, or at the very least he wasn't
- 46:39an engineer in the, in the sense that he's Wozniak was an
- 46:43engineer. And so there is this huge skill
- 46:46set to figuring out one, what are the boundaries of the
- 46:51technology? What are, what are the limits of
- 46:52the technology #2 how do you best percent that technology #3
- 46:58how do you capture people's attention and distribute it?
- 47:01And so I think Steve Jobs was really good at these things.
- 47:04And then he quote, UN quote, used the team of agents, the
- 47:08people that he recruited. Well, as human beings, right?
- 47:11Yeah. But you get to kind of see where
- 47:14this is going, right? There will likely be a bunch of
- 47:18Steve Jobs that figured out that trifecta and are offloading the
- 47:23hard engineering work to agents. But also there there might be
- 47:28teams constituted almost entirely of Steve Jobs people
- 47:32like, we're all tastemakers in this room.
- 47:35We all bounce ideas around and like what kind of products we
- 47:38want to see in the world. We do a bunch of research.
- 47:40We talk to customers, but we're basically like, I would say it's
- 47:45almost like the return of the product manager.
- 47:48I think you're afraid that iconic moment in culture of the
- 47:53PS by the pool. The intersection of engineers
- 47:57resenting product managers and general Twitter loathing of
- 48:00women, Yes. I'll tell you, like, I don't
- 48:02think people will have seen this coming, right?
- 48:06But there is a return to, you know, well, do you need more
- 48:10than the PM with the person sort of like understanding market
- 48:15customers requirements, E etcetera.
- 48:18And so it's almost like a pendulum swing back and some on
- 48:21some levels at least. But yeah, I think the, the, the
- 48:24reality is is likely to be a lot more nuanced than any of this
- 48:27prediction. But my, my, my personal
- 48:30predictions in five years, this whole industry does look
- 48:34extremely different, Extremely, extremely different. 16 month
- 48:37prediction Six months from now, what will be different?
- 48:41Six months. I do think there's going to be
- 48:44maybe even sooner than this. I think people are
- 48:46underestimating GBT 5. I think they're judging it by,
- 48:50again, judging a model is becoming an important industry,
- 48:55right? There's companies like LMRE.
- 48:58Takes. I'm like, I really need to sit
- 48:59with a model and sort of really understand.
- 49:02And that's an emergent thing. I I don't remember technologies
- 49:04where like you really had to sit with them for like like aiding
- 49:08to know them like a person. Right.
- 49:11It's like, yeah, you want to assess their soul, like on the
- 49:13first meeting, you know, exactly.
- 49:15We have to sit with it and sort of see what it can do.
- 49:18It's like it's like you're 306090 day plan for a new hire
- 49:24or a new executive like you have to treat a model.
- 49:26Underestimating GPD files? Yeah, I think.
- 49:28I think people are. I really think people are.
- 49:31I also think you are not realizing that this vertical
- 49:34agents, there's probably going to be some agents, maybe some
- 49:38like agents for video editing or creating video or creating ads
- 49:43or you know, I can't tell you exactly what.
- 49:45There's going to be one that we're all going to be talking
- 49:47about. Interesting.
- 49:48Yeah, Yeah, we need to break out Chachi.
- 49:50BT has so much defined the use case.
- 49:52We need sort of a vertical niche where it's like it's going to
- 49:55solve this problem, do this thing calendar.
- 49:57Something along this line like it seems so odd but how do we
- 50:01not do it? As soon as Chachi BT came out,
- 50:03we didn't do it because we didn't have the infrastructure,
- 50:06we didn't have the best practices, we didn't have the AI
- 50:09gateways, we didn't have the agent frameworks.
- 50:12We built a lot of this stuff. And so there's going to be an
- 50:16obvious in retrospect application with a huge user
- 50:21base. That's my.
- 50:22Perspective right now that Gmail is not solving look that the
- 50:25model is not in my e-mail yet To me is the, I mean superhuman
- 50:28wants to do it, but to me like Gmail, it's like sitting there.
- 50:31There's a lot of risk. Obviously I would be a bad
- 50:34reporter if I didn't ask. You can dodge or not.
- 50:37The information says you're getting approaches at 9 billion.
- 50:41Do you have anything to say about that?
- 50:44No comments. No comments, speculation.
- 50:46All right, Guillermo, thank you so much for coming on the show.
- 50:49We have to do this again. This was a lot of fun.
- 50:51Hey Gary, always play.