Latest / Newcomer Pod / The Artificial Intelligence Startup Draft
Transcript
- 0:10Welcome to the Cerebral Valley Podcast on newcomer I'm Back
- 0:14with Max Child and James Wilsterman, my Cerebral Valley
- 0:17Co host and the Volley Co founders.
- 0:20We are very excited about this episode.
- 0:22The meat of it will be a draft of the most valuable AI
- 0:29companies we get. It is like, it's as close as I
- 0:32think in Silicon Valley you can get to watching sports hopefully
- 0:35or fan of. Sports.
- 0:37Sports debate TV show. Yeah, exactly.
- 0:42Yeah, I I overcomplicated it to make sure I got the.
- 0:47I won't spoil it anyway, you know we took it very seriously
- 0:51and even if all our listeners forget about it, I am sure.
- 0:55James, what was the time window? It's five, is it?
- 0:58Five years, Yeah. In five years we'll all be
- 1:02checking to see see how we did. But before that Cerebral Valley,
- 1:07the conference is on Wednesday. And you know, obviously for the
- 1:13people going in person, you, you get why you're curious.
- 1:16It's like OK, But then for everybody else, the videos from
- 1:20the conference will be posted on YouTube and in the
- 1:23newsletternewcomer.co. So this conference will be
- 1:27brought to you digitally. We have not just hyped up a an
- 1:29event you will not get to consume, you will get to see the
- 1:33talks. So look forward to that.
- 1:35I think usually we put some of the best ones in the podcast
- 1:39feed and we just wanted to sort of take stock going into the
- 1:44conference sort of big themes that that are going to be on our
- 1:49mind and then what's changed from the first Cerebral Valley
- 1:55in March. So yeah, let's get to it.
- 1:58All right. I said you guys could could go
- 2:00for themes last, I mean go for themes 1st and I would play
- 2:05clean up so. James, why don't you go still
- 2:08thinking about my themes? Yeah, I have noticed a theme
- 2:13kind of gaining some heat over the last few weeks on Twitter.
- 2:18Slash X of Kind of this big debate around open source
- 2:23versus. Close source in terms of safety,
- 2:28right. I think this ties in with the
- 2:30executive order put out by the Biden administration and kind of
- 2:36a raging debate I guess or simmering debate of whether
- 2:40these larger close source language model companies are in
- 2:45the business of regulatory capture through their efforts to
- 2:50speak around model safety and get.
- 2:55Get the Biden administration involved or whether that comes
- 2:59from a true concern and how that impacts companies who are
- 3:05putting out open source models for the community.
- 3:07OK, I'm going to jump on that. Open source versus closed
- 3:10source. I I think that's probably the
- 3:11top in my mind. And in some ways I'd almost
- 3:14frame it as like I think the first conference ended up having
- 3:18more of the pro open source people.
- 3:22You know we ended with Hugging face and Replet.
- 3:25Both of them are very pro open source.
- 3:29I mean we have data bricks back they're pretty pro open source.
- 3:33Whereas this conference we're ending the day with Vinod Khosla
- 3:38who has been somewhat critical of he he wants more sort of
- 3:43control and regulation. We'll we'll get the exact nuance
- 3:46of his open source take and then Mustafa Suleiman at inflection
- 3:51and the D Mine Co founder who's going to end the day has been
- 3:54super supportive of regulation and somewhat apprehensive about
- 3:58open source. So I think this conference in
- 4:00some ways will be the response to the really pro open source of
- 4:04the first. Of course I I think open source
- 4:06is very popular in the not open AI world.
- 4:09So I wouldn't be surprised still if probably the majority of our
- 4:12panelists are still pro open source or yeah, what do you guys
- 4:15think? I guess one thing I wanted to
- 4:18ask you guys on this topic is. Last time at the conference, a
- 4:23lot of discussion was about the letter that just came out, with
- 4:26a lot of signatories around taking a pause on AI
- 4:31development. A modest facility had signed it.
- 4:34Totally perplexingly, yeah. Yeah, which, which was?
- 4:38Confusing to a lot of people. And I guess my question is, at
- 4:43the time, it felt like there was a lot of heat on companies even
- 4:46like Open AI and Sam Altman. For being cavalier with their
- 4:52development of AI and causing, you know, an AI race and putting
- 4:55out ChatGPT. And you know, before the bigger,
- 4:59you know, more cautious companies like Google had
- 5:01released their AIS, right? And I really think in like 6
- 5:05months that has just totally changed where you know now that
- 5:08you have this open source community putting out models.
- 5:11Like Open AI is able to like look like the responsible
- 5:14stewards of AI going to the White House and you know,
- 5:18getting regulations in place. Just curious if you guys agree
- 5:22with that assessment? I do think a lot of the safety
- 5:26talk has died down. I guess the executive order is
- 5:30maybe a contrary example of that.
- 5:33Although I you know, maybe the White House timeline is like six
- 5:35months behind everyone else's timeline, but.
- 5:39It does seem like people are a lot less hung up on like is the
- 5:44AI going to kill everyone and sort of the tech industry and
- 5:47then I guess simultaneously we got this massive executive order
- 5:50banning or not banning but regulating models over a certain
- 5:53size etcetera, etcetera. So it's it is interesting to
- 5:58think about if there's going to be this sort of waves and and
- 6:02troughs or peaks and troughs of the the hype cycle around safety
- 6:06and I feel like. Burn a little bit of a through
- 6:08inside the industry right now, but six months from now that
- 6:11could change again. Just got to keep having
- 6:13conferences in March. GPD 4 was like, you know, a
- 6:16shockwave and it had come over 3.5 and I think it felt like,
- 6:21man, if this keeps going like we're going to have generalized
- 6:24AI tomorrow and then I nobody is really caught up with them,
- 6:27right? There are open source models and
- 6:30people can say, oh, this is better than open AI on cost.
- 6:35But I don't think anyone even in specialized use cases has really
- 6:39had anything as smart. So then there's just a little
- 6:41less fear of like, Oh my God, it's becoming like.
- 6:44So you know, that is just not sort of the pace hasn't been so
- 6:49out of control. And I think that's part of what
- 6:51has cooled things down. But I sort of, I mean I think
- 6:54the binding executive order is like a huge deal and shows that
- 6:57there's very active sort of government interference and like
- 7:00Europe has been talking about things.
- 7:03So I don't know, I think this question of cracking down on
- 7:07open source is very real. But yeah, but the open source
- 7:10community has certainly been vibrant and and you know, and
- 7:14the other piece of that is that firms like Andreessen and others
- 7:17have been very loud about sort of positioning themselves as
- 7:20super pro open source. Yeah, interesting.
- 7:23Next theme. All right, I have a theme for
- 7:25you guys to think about here. I think one of.
- 7:30The all time maybe the number one all time Silicon Valley
- 7:33investing philosophy for all categories of software is you
- 7:38know we like to invest in companies that sell picks and
- 7:41shovels, not the companies that dig for gold, right.
- 7:43We like to, we like to fund the tool makers, not the tool users
- 7:47in many cases. And I think what's interesting
- 7:51just in the last six months between the previous conference
- 7:55and the current conference is. The the sort of what tools were
- 8:00in vogue 6-7 months ago and then what tools are invoked today.
- 8:04And then the sort of cherry on top of that is Open AI itself,
- 8:09you know, very, very recently announced this kind of huge wave
- 8:13of first party tools improvements whether it's you
- 8:17know, retrieval improvements, context increases, speed, cost.
- 8:22Really seeing anyone able to make their own GPT a character
- 8:26to kind of go after the character AIS of the world.
- 8:28Like they really came hard on like a bunch of tools categories
- 8:32that I would say 6-7 months ago were being funded as like pretty
- 8:34compelling stand alone companies.
- 8:36And I think it's sort of a A2 way question which is 1.
- 8:42Is tools investment still good idea in an AI world like or is
- 8:46open AI going to eat this all alive and then I guess 2 like?
- 8:50You know who who is, who is still exciting in the tool space
- 8:54and and who do we think has like a strong foundation for the
- 8:56future. Yeah, I mean I think that
- 8:59question of will Open AI just eat everything applies in every
- 9:04part of the. Yeah, I mean that was a theme of
- 9:07the first conference definitely, just like how how relevant is
- 9:11everybody except Open AI and and sort of an ongoing question.
- 9:16So certainly, yeah, I think now with the developer conference,
- 9:20this thread on tools and then I think you know, given ChatGPT is
- 9:26clearly the most popular consumer AI app.
- 9:28Well, I'd be interested. I mean, you have authentic
- 9:30conversations with venture capitalists regularly, Eric,
- 9:33right behind the scenes that don't all get published.
- 9:35Like, do you ever get some authentic expression of concern
- 9:41about tools companies that they've backed in the space that
- 9:43were like, well, maybe we. Jumped on this tools idea too
- 9:47early and it actually isn't a standalone product or you know
- 9:51behind the scenes there are definitely a lot of firms that
- 9:54are shitting on foundation model bets.
- 9:56Oh, interesting. Oh yeah, that's that's like, I
- 10:00mean they're expensive you know they're very speculative.
- 10:03Open AI is in the lead. They feel like even with open AI
- 10:07open source comes to parity very quickly.
- 10:09So the, the category I hear the most dunking on is just like
- 10:13foundation models and you'll you'll see firms that haven't
- 10:16made a lot of bets there. You know, yeah, the Valley loves
- 10:20like tools, picks and shovels. So that's not something where I
- 10:25feel like I've heard a lot of regret.
- 10:26And if anything I feel like I've been, I mean I've been talking
- 10:29like vector databases all week and stuff like that.
- 10:32And literally you're getting VCs talk about chips or getting
- 10:35companies around chips, which traditionally they've been
- 10:38scared about. But you have sort of the fact
- 10:41that NVIDIA is doing so well and then you also have the fact that
- 10:44it's sort of cool to be like an emerging tech investor at the
- 10:47moment. So I think those things
- 10:48combined, you're seeing that. So no, I I think V CS dunking on
- 10:54foundation models, but bullish about everything else supporting
- 10:57AI. Yeah.
- 11:00Are there particular company types you have in mind like do
- 11:02you think? Well, I mean, I think there's
- 11:04like obvious companies that are like.
- 11:06May be less hot than they were in the last conference, right?
- 11:08I mean like Jasper. Jasper would be an example,
- 11:11right? Stability would be a huge
- 11:13example, right? But I think there's company
- 11:14stability that's just like the unique failings of.
- 11:17That sure, sure. But there are companies that we
- 11:19talked to last time, right? You know, Lang Chain, for
- 11:21example, right. That like it's a big question as
- 11:24to where Lang Chain fits in the universe, where Open AI is just
- 11:27going to take care of all this crap, right?
- 11:29And then? You know, I think we're talking
- 11:31to some companies this week that, you know, look like they
- 11:33they could totally be awesome companies.
- 11:35And I have no earthly clue, but they're providing very
- 11:38specialized tools like you know, data cleaning or something like
- 11:41that, right. Like which again I'm excited to
- 11:43hear the pitch like of of why that's a a great idea for the
- 11:45company and then probably is. But like, you know, I would be a
- 11:49little scared that opening. I was just going to like kick
- 11:51ass in my category if I were a tools company.
- 11:55And so I just think, and I mean, I think you have to agree that
- 11:57character has to be pretty upset that Open AI basically announced
- 12:01their version of character AI tools, right?
- 12:03Like, so I'm just interested if there's any real feedback of the
- 12:08fear factor on that front. But it sounds like maybe not.
- 12:10So yeah. Do you think just jumping diving
- 12:15in on the character question like there is a theme I've
- 12:18noticed of people kind of questioning whether?
- 12:22Open AI for consumers as an end point is is really sufficient.
- 12:27So there's kind of like the opposite side of that question
- 12:30from the tooling side, right? Like will they eat all tools?
- 12:32And then there's from the consumer assistant side, is
- 12:35there a super assistant that's just ChatGPT or do you end up
- 12:38using all these characters or mini GBTS that that they are
- 12:43working to enable developers to build?
- 12:46How do you guys feel about that debate?
- 12:48Yeah, I mean, I think that's, that's the open question.
- 12:51I think it depends. It depends on the quality of
- 12:54the, you know, the foundation. Models.
- 12:57I don't know. I mean, I guess like you
- 12:58selfishly we built games, James, that are built on AI tools,
- 13:01including in some cases large language models.
- 13:04So my selfish belief is, is possible to build games on LLMS
- 13:09without open AI building the exact same thing.
- 13:12It may depend on the complexity level of the consumer app built
- 13:16on top of the tool, whereas character is sort of.
- 13:19Pretty vanilla RE, you know, re prompting or you know tweaked
- 13:23version of the core, the GPT experience.
- 13:26But you know, I think it depends on the complexity level and sort
- 13:29of how much of the stack you control and and how easy it is
- 13:32to get to your customers. I mean a phrase I've been
- 13:35hearing over and over again this week, you know, just talking to
- 13:38VCs has been retrieval augmented generation.
- 13:42Rags, Rags and I think there are a lot of companies right now
- 13:45that think we can be really good at giving GPT the sort of
- 13:51information it needs and then you know Open AI or whoever can
- 13:57sort of do the thinking based on that.
- 13:59But we're not, we're not in the business of making the best
- 14:01thinking company, but we're we're really good at figuring
- 14:04out like what it needs to consider and and it's possible
- 14:08that there are a lot of what. Kind of companies do you think
- 14:11are saying that around like I guess this is kind of.
- 14:15Pointing to data and unique data sets as kind of a mode of some
- 14:19kind, right? Like, what kinds of companies do
- 14:22you think fit that category of potentially having a true data
- 14:26mode that could be valuable, having a data mode?
- 14:31Yeah, Well, well, this isn't a data mode, but I mean, I'm one
- 14:38company as you'll hear in the second.
- 14:40Isn't it a data? Isn't it a data mode that you
- 14:43are saying that? These companies are talking to
- 14:45you saying we have well, like OK, there's this company clean
- 14:49right, that lets companies run searches, right?
- 14:52Is it a data mode? They're really like good at
- 14:54hooking companies up, companies data up to data sources that
- 14:58then connect, that use rag and then plug into something like
- 15:03Open AI. So it's not necessarily like a
- 15:06Bloomberg where it's like we have all the financial data in
- 15:08the world. It's we're really good at like
- 15:10getting the relevant data four companies into into this system
- 15:14is is 1 case. But at least at Open AI Dev Day
- 15:18this week, they seem to be going after some similar use cases of
- 15:23allowing companies and enterprise to build their own
- 15:26internal GPTS or chat GPTS based on internal company data.
- 15:31So I think that that is still some area that they seem to be
- 15:35competing in as well. I mean, you know, eventually
- 15:39there's only so much one company can do unless, unless it's like
- 15:43the, I feel like James, this is your style of suggestion
- 15:46throughout this series. Unless there's AGI and like
- 15:49every way we talk about this is wrong.
- 15:51But in any sort of traditional business sense where like you
- 15:55know, Sam Ullman, you know, has joked in the past that like, oh,
- 15:58we'll just ask open it, we'll ask Chad GBT at some point, you
- 16:02know, like what our business model should be and we'll do
- 16:04that. Clearly they didn't need to wait
- 16:06for that. They figured out some business
- 16:08models. But you know, I anyway, my my
- 16:11point is just like in a traditional, normal world that I
- 16:14expect to continue, You can't do everything And so whatever,
- 16:17they're going to have to pick their strengths and then
- 16:20everybody else can do other stuff.
- 16:24I'm going to, I'm going to give a last, a last theme that's sort
- 16:28of more in the fun sort of philosophy sort of world.
- 16:32You know, fitting into that AGI question which is just how
- 16:36seriously do people take existential risk in a, you know,
- 16:42is existential risk now is seen as like this is propaganda for
- 16:47people who want to make it 1 seem really awesome and sexy and
- 16:51amazing and can do everything. And then two people who want to
- 16:54be in this world of like shutting it down, Like how much
- 16:57is existential risk really on people's minds anymore?
- 17:01Or is it much important that we just make sure it's not racist
- 17:05and make sure, you know, deal with these like, very practical
- 17:08concerns that are obviously true now versus imagining this world
- 17:12where it's. You know, you know, like our AI
- 17:15Doom episode. You know, figuring out how to
- 17:17optimize paper clips into blowing up the.
- 17:19Earth. That's kind of alluding to my
- 17:20earlier point about I feel like Silicon Valley has stopped
- 17:23taking safety as seriously, right.
- 17:24I mean, like, I think that like, yeah, sure, the executive order
- 17:27is not an example of that, right?
- 17:29But in Silicon Valley, I think the discussion of, to your
- 17:33point, the the implied seriousness with which people
- 17:37take. Existential risk seems pretty
- 17:39low to me because otherwise people would be like blowing up
- 17:44GPU factories or whatever. I don't know.
- 17:47It's like, if you actually believe this, had a 20% chance
- 17:50of killing all of society, which a lot of people say, like how
- 17:54could you morally work in this space?
- 17:57Or at least not like, you know, be actively opposed to any of
- 18:01the development here, right? And and the only people who seem
- 18:04to really act as if they truly believe existential risk.
- 18:07Is real is basically you know, Elisa Yodkowski and the sort of
- 18:10the AI doomers on Twitter, right, who are really out there
- 18:12every day saying there was, there was an interview.
- 18:15Shout out to Logan Bartlett with his podcast.
- 18:19He interviewed Dario, the Co founder.
- 18:22Of anthropic, right? And Dario gave like it was like
- 18:2520% chance, like everything. Goes there was like 10 to 20%
- 18:28chance that was like really high I think.
- 18:31Is that like in the next 100 years?
- 18:32So And again the time frames matter right?
- 18:35Like. I think it was.
- 18:36It was reasonably soon. But yeah, I I don't know.
- 18:39I sort of think you'd have to be insane to be working in an
- 18:43industry where you truly believe that progress there had a 20%
- 18:47chance of ending all humanity. I I don't know.
- 18:49Like, unless you believe in some complicated.
- 18:52Game theory about deterrence or something.
- 18:53And back to our earlier episode about, you know, the only thing
- 18:56that can stop a good guy with an AI, or sorry, a bad guy with an
- 18:59AI is a good guy with an AII. Said it wrong.
- 19:01Who coined that line? Did I get that line?
- 19:03I don't. I can't remember but.
- 19:05But the only thing that can stop a bad guy with an AI is a good
- 19:08guy with a better AI, right? Yeah, unless you can explain
- 19:11some pretty elaborate game theory on how progress in AI by
- 19:16the good guys is necessary to stop extinction.
- 19:19It just, you know, it doesn't compute that you believe this
- 19:21stuff has 20% chance of killing us like you know, and be working
- 19:26in the industry at the level of Dario, who literally is like the
- 19:29number two most powerful person in foundation models, right?
- 19:33I mean it's just, it's nonsense. So I don't know.
- 19:36I agree it it doesn't seem like people take it that seriously
- 19:39for real. As we discussed in that episode,
- 19:42though, there's a lot of, like, very bad things that could
- 19:45happen that, you know, AI kills some of us, right?
- 19:48Not. All of us, Yeah.
- 19:49That was the that was. The that was the take away all.
- 19:52You're right, right. Maybe it doesn't.
- 19:53Like, I guess like maybe maybe people have come around to sort
- 19:58of a middle ground here where it's not existential risk, like
- 20:02extinction for humanity and more, that there's a lot of bad
- 20:05things that could happen over the next 5 to 10 years with AI
- 20:08and I think all of us. Would agree with that, right?
- 20:12You know there are bad things that could occur if the, you
- 20:16know, some bad actors got access to some really powerful sure
- 20:20models or something, right? I mean.
- 20:22I think you the fact that Cruise is not active in San Francisco,
- 20:27like even if they sort of bungled their disclosures to the
- 20:30government, I mean Cruise is a better driver than humans in San
- 20:35Francisco, right Like. They really they they.
- 20:37They broke the one rule, which is don't lie to the government,
- 20:41right? I mean, like, it's just like,
- 20:43you can be, you can be bad, you can screw up, you know?
- 20:47But don't lie to the government is like freaking like #1 on the
- 20:5110 commandments of running a company, right?
- 20:54I mean, like, you know, that's why SPF travels or whatever, You
- 20:57know, like, I mean, like like just don't lie to the
- 21:00government, like, you know? Fake it so you make it can apply
- 21:03in many scenarios, but not with the government.
- 21:05There, you can't do that with them.
- 21:08So anyway can we just go back to this for a second?
- 21:11So you're you're what are you saying Eric about existential
- 21:17risk or Max like you you guys are saying companies are not
- 21:20acting like they believe their own hype or do you agree with
- 21:24that or? I think, I think people are now
- 21:26saying calling bullshit in sort of private.
- 21:30Existential risk. And some of them don't want to
- 21:32say it publicly because they don't want to be seen as sort of
- 21:35undermining legitimate concerns. But I think a lot of them are
- 21:39just like doesn't feel close and then how do you feel that ties
- 21:44into the executive order? Does that feel to you like over
- 21:50regulation or you know, pre? Premature.
- 21:53I'll be sort of waiting until after the conference to stake
- 21:56out my position partially, just like Get.
- 22:00Get more information And also, you know I sort of shift between
- 22:06information gatherer and sometimes opinion haver.
- 22:09And I don't know. I've been sort of waiting to
- 22:11come out strong, though I do have a sort of.
- 22:13Side the last theme I'm actually interested in, and this is very
- 22:17close to home for James and I in the last couple of months, is.
- 22:21We are increasingly discovering that the cost of running any
- 22:24large language model products at consumer scale is unbelievably
- 22:29high. Like, oh, interesting.
- 22:31Yeah, is so high that it's it's very hard to come up with a
- 22:36mental model for how you could build anything on top of a GPT
- 22:40for quality model right now that would be distributed to
- 22:43consumers. You could maybe do it on 3/5.
- 22:47Especially with some recent price cuts with, you know, fine
- 22:50tuning, yadda yadda yadda. But I would say that it is
- 22:54shockingly high to run a very basic application that's GPT 4
- 22:58level quality on LLM. And I think it basically makes
- 23:02it impossible to build consumer apps on these level of models
- 23:06today without incinerating money.
- 23:09And so I think it's an interesting question as to
- 23:13whether or not a lot of these companies that have raised 100,
- 23:15two $100 million dollars. I have to imagine their unit
- 23:19economics are pretty much upside down if they're building any
- 23:21sort of consumer product. As in every time someone uses
- 23:25their product they lose more money than they make.
- 23:27So I wonder if the only outcome of this is that B2B companies
- 23:34can maybe afford this because they can charge such such
- 23:36staggering prices. And then?
- 23:39Open AI is like the only one that can do consumer stuff
- 23:41because you pretty much have to have no middleman to afford
- 23:44running these applications. OK.
- 23:46So that's a great one and I'll broaden it and just say I think
- 23:49there's an the overall meta theme of that is the rubber
- 23:53meets the road, right? There's this.
- 23:55Do people really care about existential risk anymore?
- 23:57Because it's like, oh, we're just in the practical problems.
- 24:00Like it's really expensive to run consumer apps and I've heard
- 24:04people in enterprise ask like are.
- 24:07Are big companies, customers sticking around or do they try
- 24:11the demo and say we didn't like it anymore?
- 24:14And I mean, you know, Chachi, BTI think we're going to see in
- 24:17a slide during the presentation, you know, activity syncs and
- 24:20then they have new features and goes up like what's the
- 24:23resiliency? So just this, yeah, the rubber
- 24:25mean the road, all right, We've gotten all the hype like how are
- 24:29people delivering? And now it's sort of the moment
- 24:31where it's turning into sort of more regular sort of business
- 24:35style questions. I mean, this stuff is so
- 24:39expensive to run that it's hard to imagine you would not have to
- 24:42run into regular business style questions pretty quickly.
- 24:45Now I was not in the Internet industry in 1993 when you had to
- 24:48go buy your own servers or on a website, you know, so that this
- 24:54may be analogous to that time, you know, where you had to be a
- 24:57garage band server farm to afford to run these kind of
- 25:00companies. But even that would have been
- 25:03that would have been fixed cost, right, I mean now we're talking
- 25:06about. Like we're talking erasing cost.
- 25:08More customers you get, more expensive.
- 25:09Yeah, Erasing software margins like.
- 25:13So like, I'm pretty interested to see if anyone can run a
- 25:17consumer product without just incinerating money besides Open
- 25:22AI. So.
- 25:24Stay tuned or or or use I would say.
- 25:26I would say the last caveat would just be using smaller
- 25:28models or. Yeah, using smaller smallers or
- 25:31fine tuning or something. Or GBT 4 gets 20 times cheaper
- 25:35in the next year and a half. And then this is a dumb,
- 25:39irrelevant, you know, retrospective discussion, but it
- 25:41really has to go down by like a 10X to be interesting.
- 25:44So they did three XA week ago. We need two more of those to
- 25:49really be in a good category of cost.
- 25:53Yeah, we, we should. I mean, that is a good reminder
- 25:55that we should ask money. Like, how much does this cost?
- 25:59Like, how? Yeah.
- 26:01How Where do you make money if like, you know, prices change
- 26:04one way or the other. Yeah.
- 26:05Yeah. All right.
- 26:07Next up, I really think you're going to want to listen to this,
- 26:11the draft pick. Please give us your feedback.
- 26:14I want insane like memos about like what the order strategy
- 26:19should have been, and just feel free to weigh in.
- 26:23We did, you know, I I think we explained all the particulars in
- 26:26in James, right. You talked through Yes, yeah,
- 26:28we're about to jump into the rules and there were, there was
- 26:32probably more time spent behind the scenes debating the rules
- 26:35than I preparing for the draft. I was insistent about a
- 26:38particular rule change that I mean you'll, you'll hear.
- 26:42So we had a lot of fun. Sent you Eric at newcomer.co.
- 26:48You can always send me your opinions about who's an idiot I
- 26:52hopefully. I think I'll try and post the
- 26:54list. So please forgive us that it was
- 26:57a set list. There are companies.
- 26:59I asked if I could pick companies off list as sort of a
- 27:03insider with Intel. But I was I was not allowed to
- 27:05pick off list. So I you're restricted to a list
- 27:08of companies and we didn't pick you, please hire a new marketing
- 27:12officer. I don't know.
- 27:13Sorry. Yeah, find find out why your
- 27:17your valuation was also not, you know, listed in Pitchbug, right
- 27:21'cause the list was selected from companies who are or who
- 27:26have have a recent over $100 million rounds in Pitchbug.
- 27:32There's go. Listen to that.
- 27:37Welcome. Welcome.
- 27:38All right. We are gonna move into the draft
- 27:41we this is probably the thing we're most excited about at
- 27:45least informed to do taking a page out of.
- 27:50Sports Podcast, Political Podcast Playbook and we are
- 27:54going to do a draft pick of the top AI companies.
- 28:00James, do you want to explain the metrics for who wins this
- 28:05and how we know and all of that? Here's what we're doing.
- 28:08Max, Eric and I are drafting teams of AI startups.
- 28:13The startups must have risen over $100 million.
- 28:18I wish that wasn't the case, partially because I have Intel
- 28:20on like companies that are promising, but James isn't
- 28:24letting me include random. I feel like there's, you know,
- 28:27it's a risk. They are not even valued at that
- 28:30yet, but I I, yeah, I think we have, we have 43 companies in
- 28:35the list and we're only picking five each.
- 28:39So many of these will not make it on the board.
- 28:42So I'm excited to see what happens, but this really shows,
- 28:46I think, the amount of companies who have raised ginormous rounds
- 28:51in AI. Is this 100 million plus list
- 28:54you have to have raised raised over 100 million, not valuation
- 28:57raised. It's probably 100 million
- 28:59because I I do think also they're probably ones that
- 29:01aren't. So some other criteria I used to
- 29:05kind of narrow down this list. You had to be using generative
- 29:09AI or the company had to be have generated AI or generative AI
- 29:14infra kind of core to it. I kind of excluded some
- 29:17industries like bio and healthcare, defense, silicon
- 29:22chips, robotics, driverless, or a VS and China.
- 29:26I just wanted to filter down to companies we might have a chance
- 29:30of knowing something about or being predictive about, rather
- 29:34than just predicting whether some biotech company is going to
- 29:39use AI to discover a new drug or something.
- 29:42And I also wanted companies founded before excluded,
- 29:47companies founded before 2013. So most of these companies have
- 29:51been actually founded in the last three years or something,
- 29:54but there are a few who have been founded maybe earlier.
- 29:58So that's the game and we are going to determine who goes
- 30:02first by creating a handicap auction.
- 30:06So this was, I was adamant, yeah, that we do this.
- 30:09James and Max were like, oh, you know, that's part of the
- 30:12randomness. But to me, like a huge part of
- 30:14this is how how much would you pay to get to pick?
- 30:19Open AI Because I, I, I honestly think we're all going to pick
- 30:22Open AI. We should talk about what the,
- 30:24what the, what the grading criteria is.
- 30:27Yeah, here. So the way we are drafting these
- 30:31teams, we are we are trying to create a team with the highest
- 30:35total valuation on November 1st, 2028.
- 30:40And final valuations would be determined by pitch book, either
- 30:44kind of a reported M and a price or a current stock price if if
- 30:48the company is public or the last evaluation if the company
- 30:52is still going but is private and if the company is shut down,
- 30:56I think that would count as a 0 for that team.
- 31:00If if like 2 year, does it have to have raised at the price
- 31:04within the last two years or something because?
- 31:09There's whatever the last fundraise was.
- 31:11Yeah, just. But like, it might not be worth
- 31:12that anymore. One question I have is let's
- 31:15let's say there's a company on this list who has raised, who
- 31:19has raised a huge round, maybe in the Zerp era but never raises
- 31:23again, but is a zombie company in five years.
- 31:27Are we counting that at their? I think the last valuation.
- 31:31No, it needs to raise another round or do something in the
- 31:35next five years because otherwise there's a strategy of
- 31:38just like picking the most highly valued companies.
- 31:41And just betting that like nothing happens.
- 31:43That's the efficient market hype.
- 31:44Out there, I don't think. I think it's theory, current
- 31:47evaluation, other. Is the best particular of future
- 31:50value. It needs another either of a, it
- 31:52needs another fundraiser. Reporting.
- 31:53Yeah, I think so. Max, are you OK with that or.
- 31:57I mean, I don't care. I don't care.
- 31:58Yeah. Yeah, yeah.
- 31:59So whatever. It needs something, whatever,
- 32:01right? Are we agreed on that?
- 32:03Yeah, we're agreed. So if the company does not raise
- 32:07any round between now and November 1st, 2028, yeah,
- 32:12according to Pitch Book, a new round, they will count as a zero
- 32:16as well. Fair.
- 32:18OK. Yep, brutal.
- 32:21So how are we doing the bid for 1st and 2nd place?
- 32:28Well, I think the the idea here is if you you know we're gonna
- 32:34just go down the line here and name a price that you are
- 32:38willing to subtract from your total score and then in the end
- 32:42and then determine you know the order in that in that order of
- 32:47those 3 handicaps and we will deduct those scores at the.
- 32:51End. All right, let's just do a live
- 32:52auction. Like whoever gets the highest
- 32:54price and then we stop it. OK, so then it's a coin flip.
- 32:57OK, fine. Coin flip for a second?
- 32:59Yes. OK, let me pull up my digital
- 33:02coin flipper. OK, All right.
- 33:06And then it's a snake draft. Yes.
- 33:08All right. Let's let's start.
- 33:10I I bid -45 billion. For open AI.
- 33:15For first for to go first. For first take, OK, I'll go
- 33:26I'll, I'll go to 50. Sure, I'll go.
- 33:30I'll go 55. I'll go. 60 60 billion -60
- 33:37billion. How is opening eye going to have
- 33:40an exit? I just am confused by this.
- 33:42It does they they don't need to have an exit, they just need to
- 33:45have a fundraise. OK.
- 33:46I'll go. I'll go to 70. 75 OK, that's
- 33:58fine. Max, you're done.
- 33:59I'm done. I'm out.
- 34:00You got it at 75, all right. 1st pick You have a lunatic open AI.
- 34:05OK, bad pick. Bad pick now I.
- 34:11Imagine I did all that to get like anthropic or?
- 34:13I just, I just I just think the number of companies that are
- 34:16worth like hundreds of billions of dollars on the secondary
- 34:19market is just extremely low. So you're you're basically
- 34:21discounting the net equity value to like 5 or 15 billion now or
- 34:24whatever it is. So sorry, yeah, added.
- 34:28Yeah. It's currently valued at 80 or
- 34:2990 right on secondary markets and.
- 34:31You took 7075 off that. So you're basically saying I'm,
- 34:35you know 1010 is where I my starting point and I have to get
- 34:39another 100 billion or whatever on this valuation?
- 34:41If it's, is it a trillion dollar company like?
- 34:45Yeah. Then I crushed it like.
- 34:47But are they going to be a trillion dollar company as a
- 34:49subsidiary of Microsoft? I just don't believe that.
- 34:52But that's fine. This is why it's an exciting,
- 34:54contentious pick. Yeah, great.
- 34:56OK, All right. Eric's going to go 1st, and Max
- 35:00and I are flipping for 2nd place.
- 35:02Max, name your head or tails here.
- 35:04I'll take heads. Heads.
- 35:06You're up second. OK, Eric, we know what you're
- 35:08picking. Do you want to make it official?
- 35:10Open AI Congratulations knew what he was doing.
- 35:15Smart guy. Smart guy.
- 35:17Anything you want to, Anything you want to add there, or.
- 35:20Well, I I do also think the fact I think Microsoft be an investor
- 35:24helps drive up the valuation because they're getting spend
- 35:27back on their system. There's all this revenue round
- 35:30tripping going on. I think it'll definitely raise
- 35:32another round or two. So I feel like it's super safe
- 35:35bet. I think I'm getting it if we
- 35:37know it's reported to be 90 billion.
- 35:39I'm getting it like, yeah, with. With.
- 35:41But if Microsoft buys, if Microsoft buys an additional 2%
- 35:46to go from 49% ownership to 51% ownership, that's that's like an
- 35:49exit in my opinion. Like that's an exit.
- 35:51I agree. Yeah, yeah.
- 35:52What do you mean no? They have majority ownership of
- 35:55Open AI. I mean it's it's a scenario of
- 35:57Microsoft at that point. No, it's just over then, like.
- 36:00Yeah, I think. So no, these are if there are
- 36:03deals on Open AI, they value it as more.
- 36:06It's worth more, definitely. If other people, if people are
- 36:09getting shares at a higher price.
- 36:12There there will be no shares of Open AI, though in this context
- 36:14I don't know we can we can see what actually plays.
- 36:16At it, the valuation goes up and there are people who own a lot
- 36:19of it, then yeah, it it has its own valuation.
- 36:23There's no way that you're saying, well, well, I I I guess
- 36:27the shares have to dissolve is what you're saying for it to be
- 36:29an acquisition that's kind of. Yeah, yeah.
- 36:32Once they're in Microsoft, I accept that.
- 36:35Like I don't get to ride up. The Microsoft benefit.
- 36:38Of. Even though I do think that's a
- 36:40possible outcome, but. I I just don't think they ever
- 36:42make it to whatever a trillion dollar valuation is a stand
- 36:46alone. I don't know if Sam Altman would
- 36:48sell, oh, you know, away Control to Microsoft.
- 36:51That's I don't think so either And like there's antitrust and
- 36:53like I think this is the most Microsoft is probably going to
- 36:56own. So, but I I'm not willing to
- 36:58accept like if they own 60%, I'm like no, that's fair, that's
- 37:01fair. OK.
- 37:02All right, all right. All right, Mac.
- 37:04Max. I will take data bricks at #2
- 37:11Yeah smart, pick smart I. Like I like Data Bricks, I I
- 37:15like to think that I mean the goal here is obviously to really
- 37:20get a a couple huge winners or one really huge winner, right.
- 37:24And I think that Data Bricks is obviously already a very large
- 37:27company, you know 10s of billions of dollars and I think
- 37:30there's probably many 10s more or or potentially hundreds in
- 37:33headroom there and so. Yeah, I don't know.
- 37:36I think it's a good call option on a really massive company.
- 37:38So, So I'm grabbing data bricks and also my favorite speaker
- 37:41from CV1 and Ali goes to the CEO and he's coming back for CV2.
- 37:46So I'm just a little Homer pick here.
- 37:48Smart. Yeah, this is giving me now.
- 37:50My stomach's turning. I do data bricks like I've been
- 37:52super bullish about them for the long time.
- 37:54So not to have them in my list is it hurts.
- 37:58All right, James, pick 3. All right, I've got two that
- 38:01I'm. You get two in a row, Snake
- 38:04draft. You get two in a row.
- 38:05I get two. OK, I just do both of them.
- 38:07Yes. All right.
- 38:08I am going to go with hugging face as my first pick.
- 38:13Wow, that's that's high. That's high.
- 38:15I think you could have gotten the value bet later in the later
- 38:18in the draft there. No, I might have.
- 38:20No, it sounded like Eric was going was eyeing A hugging face
- 38:23as well. Yeah, I think.
- 38:25I just, I just really think Hugging Face is going to exit it
- 38:27under 20 to Microsoft basically. So that's well that's good for
- 38:30me though that's like a pretty good lock in of 20 million of
- 38:33you. Need you need to stack those
- 38:35hundreds to win this draft, James.
- 38:37We're going up against the trillion dollar open AI here.
- 38:40So then you should only pick foundation model companies, you.
- 38:43Should pick whatever you think can go to a trillion Eric, you
- 38:45know it could be anything hugging face has a lot of room
- 38:48for I think potentially getting to that 100 billion level if
- 38:53they kind of maintain with their position in in essential you
- 38:58know position in the open open source community and they
- 39:02potentially even build tooling around this that helps other
- 39:06companies use open source in their in their enterprises.
- 39:10And yeah, I just think there's a pretty high ceiling there.
- 39:17GitHub sold for 7.5 billion. Yeah I'm just, I'm just saying
- 39:20GitHub is is everything hugging faces and more in my opinion and
- 39:23they decided to exit at the top at 7, 1/2.
- 39:26So I don't know. I mean, I I love hugging face
- 39:28for the record, we're, you know, big fans, but I don't know, not
- 39:31not sure about the pick on the upside.
- 39:33And there are forum at the end of the day like.
- 39:36Yeah. All right, James, second pick,
- 39:37hosting. I I mean, we love them.
- 39:38But for my second pick, I am going with Anthropic.
- 39:44Huge second runner in the foundation.
- 39:47Model Wars, I would argue, just recently struck a pretty major
- 39:52partnership with Amazon and has a lot of the same people who
- 40:00were originally working at Open AI.
- 40:02So I think, yeah, and I've heard really, you know, positive
- 40:06things from developers about Claude in general.
- 40:09So yeah, I'm really excited about Anthropic.
- 40:12OK, this is a tough pick. I'm going to, I'm going to skip
- 40:17through the really high fundraisers here for my second
- 40:20pick, go a little more deviant. I'm going to take Pine Cone, the
- 40:25database company. I think that what I'm hearing
- 40:30from my little birdies, Little birdies in the, in the AI
- 40:34communities, Pine Cone is the the database of choice for AI
- 40:40developers of all stripes. And I think if you look at the
- 40:44history of Silicon Valley, you see really strong exits from a
- 40:48lot of different database companies.
- 40:50So yeah, I mean I I like Pine Cone as I don't know, I mean
- 40:54this would be a really good pick if I could get you know
- 40:57discounted for the fact that their valuation is way lower
- 40:59than most of the companies on this list.
- 41:00But regardless, I'll take the upside.
- 41:03I, you know, as a reporter does went went to a VC for advice
- 41:07before this. I got one because I asked at the
- 41:09last minute but he said pine cone not there on his list.
- 41:13He only gave me his five because I think her #3 enterprises are
- 41:18wasting tons of time building toy prototypes that will
- 41:21ultimately be replaced. All I all I care about is just
- 41:26the crazy upside case and I think that's that's what I'm
- 41:29going for here, so all. Right.
- 41:30I'm gonna do I get 2 now? I don't know.
- 41:32Do you get 2 now? Do you get to?
- 41:33Yeah, you get 2. Now you get to swing.
- 41:34Back. I'm like, nervous.
- 41:36I'm gonna take inflection. Please explain.
- 41:46You got to defend your picks here, Yeah?
- 41:52With the idea one sort of top talent obviously like DeepMind
- 41:57Co founder Mustafa Suleiman. I'm not picking him using
- 42:00speaking but you know I I do think super legit company.
- 42:06I mean obviously you're betting a lot on the future.
- 42:09I do think you know if it's go for big swings, huge foundation
- 42:14model potential is key smart technology I think.
- 42:21I think also just like give it, I feel like there's a floor and
- 42:24that it's also like a potential like acquisition to just like
- 42:27get some big tech company capable worst case.
- 42:33So yeah, I'm going inflection for #2, OK.
- 42:39And this is a tough one. All right.
- 42:42I'm taking character. AII Just feel oh man.
- 42:45I feel like that bad pick bad. Pick.
- 42:49You would have taken it, right? Yeah.
- 42:51Bad. You would have taken it.
- 42:53Or the pig bad. No, no, no.
- 42:54Bad pick is just a way I'd joke about every pick being bad.
- 42:57Sorry, would you have taken character like I never would
- 42:59have, but it sounds like James would have so.
- 43:01I think I would have taken it, yes.
- 43:04I'm going on a foundation model, forward strategy and I I think
- 43:09character, you know, has. Perhaps a clearer path to a
- 43:13product than most, I think. I mean they have a very talented
- 43:17CEOI mean Theresa Horowitz is going big on them.
- 43:21So what they're going to get lots of follow on funding
- 43:24rounds. Yeah, we'll see.
- 43:26I don't know. I you know something I I still
- 43:31think like the, you know, right now the bot sort of relationship
- 43:35bot thing is a little overhyped, but.
- 43:38You know, there's time. What do you think about
- 43:40character as it is today? What is your?
- 43:44Do you agree with The Take that it's all sex chat or or?
- 43:47Not, I don't, I don't know, I'm not like an expert on the
- 43:49product. But to me there there's sort of
- 43:52you got a dual bet in that they're super hardcore building
- 43:55their own foundation model deep on technology and they're
- 43:59actually have a product that's out there that people are
- 44:02finding fun ways to use. And like, I mean, you know, Mark
- 44:05Zuckerberg, that's sort of with hot or not so like anything.
- 44:08That that they have the sort of instinct to say like let's get
- 44:11users, that that to me is like a good, good muscle.
- 44:15And like, even if this first iteration isn't like the one you
- 44:18want the New York Times to be writing about, I think I think
- 44:21it's a good sign. I think it's a good sort of out
- 44:25of the money. You guys would not have.
- 44:26Picked. Inflection though?
- 44:29No, I I think. Inflection's a reasonable I like
- 44:31your picks. I like your picks.
- 44:32I don't. I honestly don't really know
- 44:34what these foundation models, which is why I'm going to pick a
- 44:36foundation model next, baby. Yeah.
- 44:42I mean, I literally think I could be choosing foundation
- 44:44models. Pass open AI and anthropic as a
- 44:48a monkey. Throwing darts at the proverbial
- 44:50board would do better than me here, but I'll just go with Co
- 44:53here. I'll just take the market
- 44:54deciding it's worth a lot. It seems like it's the number 3
- 44:57#4 best foundation model. Who the F knows is my defense of
- 45:02this pick, but I got to grab a foundation.
- 45:05Everybody wants to be in business, so I feel like they're
- 45:08going to face a ton of competition.
- 45:10They were early, sure. Yeah, I'm.
- 45:13I'm not. I'm, I'm.
- 45:14I'm freely willing to admit this with them.
- 45:15I'm I'm freely willing to admit this is my lowest conviction
- 45:18pick here. But I just feel like I got to
- 45:19have a foundation model, and that to me is the best pick on
- 45:21the board. So so be it.
- 45:23Onward and upward. Maybe I'll get lucky.
- 45:25OK, I'm. I'm ready with my 2 picks.
- 45:27All right? Please fire away.
- 45:28All right. 1st pick go go on with another Foundation model of
- 45:33course. Got to have one.
- 45:35I will say you can see why VCs are investing big in foundation
- 45:40models, because if the goal is to go for the fences, you're
- 45:43like you know, foundation models are are are you know big swings.
- 45:48And so, like nobody here is really so far trying to
- 45:51accumulate like a but like I could name some sure things but
- 45:57then. You know you want like a swing
- 45:59anyway, James. I'm going with AI 21 Labs, which
- 46:04is a Israeli foundation model company.
- 46:09Solid. Just just announced $155,000,000
- 46:13raise in August at a $1.4 billion valuation and.
- 46:21The. Participants included Google and
- 46:25NVIDIA. They create foundation models
- 46:29particularly targeted at writing, and they are known for
- 46:34their cutting edge Jurassic models.
- 46:37So yeah, I'm impressed with the founding team and the product.
- 46:41Yeah. And they they've got some, I
- 46:42think they have a partnership maybe with Amazon, like I've
- 46:44seen them come up, but it feels like they're under the the radar
- 46:48a little bit. I think it's a little bit of an
- 46:49under the radar pick, but a solid that.
- 46:53Makes sense? Foundation models, I mean I
- 46:55would say that is, you know, if if they are Israeli and they're
- 47:01valued at 1.4, probably in the US they'd be valued at 14
- 47:04billion. So I think, yeah, exactly like
- 47:09you know that that seems like a good, good, good cost of living
- 47:13adjustment right there. All right and with my next pick
- 47:17I am taking Replet. I'm Jod.
- 47:21Yeah, another CVCVAI alumni Alumnus if you will.
- 47:28So replica. You have ugly Face and replica.
- 47:31They're on the same panel together.
- 47:32You're just a you're just a big CVAI chapter here.
- 47:36I love it. I like to meet the founders and
- 47:38get to know them. Practicing for In the end, it's
- 47:46really, it's really just a bet on the founders.
- 47:48It's really about the people. It's about the people.
- 47:49It's really about the founders. What do you think about this
- 47:52stuff? Yeah, Founder market fit, if you
- 47:54will. All right, What?
- 47:55What's your? I'm definitely as founder.
- 47:57Twitter are are repla and hogging face like in the like
- 48:01trillion dollar case. They're like fighting with each
- 48:03other right? They're both like in the coding.
- 48:05World, I don't. I don't agree with that.
- 48:06I think that repla, the trillion dollar case here is that
- 48:10software engineering. Changes dramatically due to AI
- 48:15and we enter this world of everyone's an engineer, you
- 48:20you're using a cloud based IDE, you can use AI really
- 48:25aggressively as you're coding and you don't even need to
- 48:28really learn how to code. And I think they're the ones who
- 48:30are furthest along at kind of building that vision that I, I
- 48:34do believe is possible with AI that was that was not on my
- 48:37draft board, all right. Yeah, I don't think I would have
- 48:41picked it. Yeah, I'm going to go with
- 48:44another Homer Pig. I will be taking Modular,
- 48:48another founder. I will be interviewing on stage
- 48:51at CVA I2 for the for the back story here.
- 48:54For the back story here. Chris Lattner, who Co founded
- 48:57Modular, literally invented the Swift programming language that
- 49:01is on every iPhone, every Mac, every iPad in the world.
- 49:05Probably one of the top five to 10 most used programming
- 49:08languages ever invented. So he's obviously a beast.
- 49:12He has shown tremendous facility in creating really low level
- 49:17technical products that are used by literally hundreds of
- 49:20millions of people, if not billions.
- 49:22Before that he created a tool called Low Level Virtual
- 49:24Machines that is also basically everywhere.
- 49:26So I'm just saying Chris Lattner, he gets WS and he is
- 49:31trying to reinvent the programming language for AI and
- 49:34replace Python with a new language called the Mojo that
- 49:37modular is invented. And so, look, I frankly don't
- 49:40know what the business model is here or where we're going with
- 49:43this part, but you know, Chris Ladner puts points on the
- 49:46scoreboard. So that's what I'm hoping he'll
- 49:48do for my team here. So modular.
- 49:49AII am breathing like a huge sigh of relief because I took
- 49:54character instead of this company thinking the character
- 49:56was going to be a more. Competitive pick and I think you
- 50:01were right, proved true and this company did not get scooped up.
- 50:05I am taking what I would have actually probably put as my
- 50:09third pick, but strategically dropped to 4 clean which is sort
- 50:14of doing corporate search. Sequoia Back Company,
- 50:17Lightspeed, Ravi, the key guy at Lightspeed is on it, and Mamoon
- 50:23Hamid at Kleiner Perkins is an investor in this company.
- 50:26So I just feel like it's one of these, Like, I don't know.
- 50:29What do they all know? So you were really like a
- 50:32venture capital fundamentalist here.
- 50:35Like, yeah, like, you know, Eric's like, have you ever heard
- 50:40of that fallacy of appealing to authority?
- 50:42Well, I go the other way and say it's the way.
- 50:45Well. Like, you know, like you should
- 50:46always appeal to authority. I mean what we don't have the
- 50:50financials, what are you doing? You're looking for a signal
- 50:52like. I mean what is better signal
- 50:54than like? Sequoia, Mamoon and Ravi like I,
- 50:57you know I don't know what it is.
- 50:59I think they're good VCs and like I I don't have metrics So
- 51:02what am I going And honestly my VC guru that I went to here's
- 51:06his case for glean semantic search over business data is the
- 51:10biggest use case. This opportunity overshadows all
- 51:14the in house builds people are using link chain etcetera for.
- 51:17So instead of instead of cobbling it together yourself.
- 51:22Use glean like I, I, I buy that. Wasn't, wasn't this the Tanium
- 51:26pitch, which also like you know, is sort of, Oh yeah, laying
- 51:29around? You know you should.
- 51:30You need to look into the eyes of the Co founders, right?
- 51:33Like, are you gonna actually help your investors make money
- 51:35out of this company anyway? OK, well there you go.
- 51:40OK. So now I have a last one, This
- 51:43one I'm I'm debating between like product that's solid,
- 51:47protect my sort of minimum versus like upside.
- 51:51Potential man, this is like, I I have empathy for VCs.
- 51:56It's like, should I go crazy, like hoping to dream, you know,
- 52:02or should I go sort of like product that people actually use
- 52:07today. I'd like to think you would do
- 52:09more than 10 minutes of research, which is what we've
- 52:11done on this. I don't know.
- 52:13Did do VCs do that? I don't know though.
- 52:16They get, they get the financials.
- 52:17We don't get any. I guess some of these companies
- 52:19haven't like you. You already did what most VCs
- 52:23do, which is as their VC. Friends who they should invest
- 52:26in, just go with that. It works.
- 52:31Lots of people, a lot of money doing that.
- 52:35All right, I. I'm going for the the hope and
- 52:40dream of the entire European continent.
- 52:43Mistral AI Oh, my gosh. I thought you were going to say
- 52:47something else, but yeah. Yeah.
- 52:49Vladimir Zelensky. All right.
- 52:52Mistral is, you know, super hypey.
- 52:55I think Lightspeed is an investor European, you know,
- 53:00they they're very early, super, super hypey, but you know.
- 53:07Smart, smart founders. I'm really enjoying this draft
- 53:10because I really would not have picked almost any of the
- 53:12companies you guys picked. And I feel interesting.
- 53:14That must apply to me too. So yeah, we can review each
- 53:17other's teams at the end here. I know who these.
- 53:19Six would be. So I'll tell you at the end I'm
- 53:21gonna, I'm gonna I'm I am also going to go hoping to dream here
- 53:25a little bit more. So I obviously read you know
- 53:30read and researched all the founders of these companies and
- 53:33I was very taken by this founder's vision of self
- 53:37improving foundation models. So the idea of recursive
- 53:42improvement of a foundation model and so that was the pitch
- 53:46of imbue our friend Ken who on a hope and a dream and I assume
- 53:53some really solid demos has raised $234 million.
- 53:57So so but I think if you're going for true take off trillion
- 54:03dollar companies as it were the first person to figure out how
- 54:07to make AI recursively self improving Will has a good shot
- 54:10at being a trillion dollar company.
- 54:12So I'll, I'll, I'll put some some chips on Kenjon and see
- 54:15what happens. Yeah.
- 54:16I think they're very bullish on like agents, right, I mean,
- 54:18right and self improving agents as well which I thought was
- 54:21exciting. So I'll throw, I'll throw $5
- 54:25into that $234 million pot and help them get there.
- 54:29All right, James, Final pick, final pick.
- 54:32All right. With my final pick, I am taking
- 54:35adept, interesting, very highly valued cerebral values.
- 54:42CV alum, Yeah. Obviously super high valued and
- 54:46highly funded. As far as I can tell, they've
- 54:50raised, yeah, close to half a billion dollars, right?
- 54:53Woo Hoo, Hoo Hoo. Essentially they are they are
- 54:56trying to create. A new way to interact, Have AIS
- 55:03interact with computers. Essentially allow creating
- 55:06agents that can control computers.
- 55:08I don't even know I I put this as a copilot I guess, but I
- 55:12don't even know if this is like this.
- 55:15Seems like a brand new type of concept.
- 55:19You know, creating a agents, AI agents that can interact with
- 55:22computers on your behalf. It's a pretty science fictiony
- 55:26idea, so you know, could be a. Huge flap.
- 55:31Or it could change the way you know the world.
- 55:33Works founders have left in droves.
- 55:35I feel like lots of internal turmoil.
- 55:39I feel like. It was hiked earlier other found
- 55:43you know there there was a founder break up at the company
- 55:47which which is certainly a negative sign but little scary
- 55:52but but you never know stop off from from the ashes rises the
- 55:55phoenix. I mean the.
- 55:58The CEO is, you know, David Luan, a former I think VP at
- 56:02Open AI. Certainly is there sort of early
- 56:05days, super smart guy. But yeah, can't.
- 56:09Can't keep the original team together.
- 56:12Interesting, I just think. I just I think working on AI
- 56:15agents in some capacity that can like to your earlier point about
- 56:20self self improving AI and you know reinforcement or maybe Max
- 56:26you were talking about that. I think there's something there
- 56:29that could unlock a lot of value, so I'm excited that
- 56:32they're working in that area. Sure, got it.
- 56:35My zoom out I'm I'm most stressed that I should have just
- 56:41hoped for a second and gotten data bricks for.
- 56:44Cheap because I do think next you got a great deal on this
- 56:46Section 1 there, yeah? And data bricks is like a pretty
- 56:49sure thing, almost more sure than opening eye.
- 56:52I feel like the upside. Our producer says he's picking
- 56:58the script because we're editing the podcast with that and also
- 57:00that was one I was seriously considering for five working
- 57:04product, but I didn't know how much technical under like what
- 57:08their level of. Heck, sophistication was so I
- 57:11didn't pick it. What?
- 57:11What are, what are other ones on the board, left on the board,
- 57:14Anything you guys are? Surprised, notable.
- 57:16I mean, Jasper obviously is like, everybody's afraid of,
- 57:19yeah, rapper. I feel like Snorkel has gotten a
- 57:22lot of hype with like internal data cleaning.
- 57:25There's companies. That's what I've heard.
- 57:26Yeah. Yeah, I they're they're always
- 57:28bouncing around when I'm hearing.
- 57:31Writer raised a bunch of money at 500,000,000, so I'm
- 57:34interested to see sort of there was sort of a good price.
- 57:38Lots of money back. Lots of money.
- 57:40Good price for investors. It's it's interesting, humane.
- 57:43Which makes that little pin that's supposed to replace your
- 57:45iPhone that you wear on your jacket or whatever that you talk
- 57:48to. Nobody put that on the board.
- 57:51That seems like a a tough, tough sell Hard, but.
- 57:55Hardware is hard. I just hardware is very hard.
- 57:57Yeah. Yeah.
- 57:58They just launched it at Paris Fashion Week, which is odd.
- 58:01I don't. Know I mean scale you know was
- 58:03making bank off of. Growing humans and self driving
- 58:06car problems and now they're trying to make bank at human
- 58:10reinforcement learning for foundation models.
- 58:13But it's just hard to believe, you know, it's humans.
- 58:16Like they just don't seem like that.
- 58:19You want? We don't need.
- 58:20We don't need them anymore. We.
- 58:21Don't need. I know I want a big like don't
- 58:23need this middle ground thing. I mean stability not being in
- 58:27the discussion, you know compared to last six months ago
- 58:30is is quite interesting. Obviously there's been, well,
- 58:33first of all, there's been a ton of competition and you know, Mid
- 58:36Journey has gotten super good and all that stuff.
- 58:37And then secondly, yeah, obviously there was a story
- 58:39about the founder making a bunch of stuff up, which was
- 58:41interesting. I think the journey is was left
- 58:45off the board here because they haven't taken VC funding by the
- 58:48way, but they could be a super valuable.
- 58:51Yeah, I know hard to maybe never get a valuation into the cash
- 58:53flow business, but mid journey man I would definitely have
- 58:56taken that before Mistral I think interesting the I mean.
- 59:01Yeah, I think weights and biases really good company.
- 59:04But I'm just like, what sort of, is it really going to gobble the
- 59:07world? Runway Love the founder.
- 59:10Like he's so charming. But, you know, I feel like
- 59:14maybe, you know, the narrative there is like, does Adobe wish
- 59:18they could have bought Runway for like way less than it paid
- 59:21for Figma? You know, when Runway is sort of
- 59:23the relevant company now? Yeah, I mean, lots of great
- 59:27companies. I wouldn't write off most of
- 59:30them. I don't even know what go
- 59:32student is. I didn't look at him.
- 59:33That's the most high fund rate. Did anybody research Go Student?
- 59:37I researched it a bit. So my understanding is they're a
- 59:41European company who kind of got started in this AI in this
- 59:46tutoring space and raised a lot of money and now is building LMS
- 59:50to do tutoring. But I was a little skeptical of,
- 59:53like, whether that was a, you know, current product or not or
- 59:57prototype. Yeah.
- 59:59So that's why I kind of ruled them out.
- 1:00:00My last point would just be, I was really struck in doing the
- 1:00:02research how unbelievably difficult it is to even tell
- 1:00:05what the products of a lot of these companies are or how
- 1:00:08they're differentiated. And I mean, you could really
- 1:00:10just like throw buzzwords like large model learning like
- 1:00:15reinforcement something something internal data,
- 1:00:18external data, whatever. It just spit out a new Tech
- 1:00:21Crunch article with like a fake company named him every day and
- 1:00:24I would find them very similar to a lot of the research I did
- 1:00:27for this. So it's.
- 1:00:29I mean poolside, you know, like I've written about, it's for
- 1:00:31code. I mean partially it's an
- 1:00:33earlier, it's like earlier on the foundation model journey.
- 1:00:36But yeah, there are a lot where it's like, man, it is.
- 1:00:38It is remarkable how many similar ish like either Data or
- 1:00:43Foundation model or something about enterprise plus LL Ms.
- 1:00:47companies there are that are in this category.
- 1:00:49It's it's amazing. To me there's my my zoom out on
- 1:00:53my view is that like Max, I mean a lot will fall on data bricks
- 1:00:59of course. And that anthropic could be the
- 1:01:02people plausibly think anthropic could beat open AI.
- 1:01:05In which case, like, James got a great pick discounted open
- 1:01:10anthropic. Anthropic is the only one I like
- 1:01:12from James. Yeah, really.
- 1:01:15Anthropic is when I'm. I want.
- 1:01:17I like your, I like your I like your Israeli one.
- 1:01:20AI 21 also that's kind of interesting.
- 1:01:22Yeah. What?
- 1:01:22Which ones do you like from Eric?
- 1:01:24I guess the, I mean opening eye obviously is untouchable, but I
- 1:01:27think I I think inflection is also quite strong.
- 1:01:31The rest, I'm not so sure. Yeah, character.
- 1:01:33Man. And whereas your your team's
- 1:01:36just rock solid, my my team is Immaculate.
- 1:01:39No, I would say other than databrakes, I have no sure
- 1:01:41things but I really took I took call options on hopefully giant
- 1:01:45upside. Like I could see my team having,
- 1:01:47I could see my team having three zeros on it.
- 1:01:49But as long as the two are big, I'm, I'm, I'm going for the win
- 1:01:53that way, yeah. We we really had to do a
- 1:01:57lightning, a speed run of of learning VCs portfolio strategy
- 1:02:02here. We we do not have.
- 1:02:04The date, you know any we don't get to talk to anybody.
- 1:02:07Weird is severe. This is not real VC.
- 1:02:09We were playing basically with blindfolds on, but it was very
- 1:02:12enjoyable conversation piece. It's kind of funny that we had
- 1:02:15to like pick companies that could return the fund.
- 1:02:17You know, they all have to. Like, yeah, runs, right.
- 1:02:21Yeah. All right.
- 1:02:22Very excited to see you in five years.
- 1:02:24Hopefully we'll follow up and like some of us will look like
- 1:02:27maybe all of us will look like idiots in this space will be
- 1:02:30wildly overhyped and maybe my, you know, -75 billion will like
- 1:02:34swamp the total value and then we can just.
- 1:02:38I don't think that's likely, but as a closer or do you just want
- 1:02:41to read off everyone's team? Oh yeah, For people not
- 1:02:44watching, the teams are. Team Eric with a negative $75
- 1:02:49billion handicap that I paid to go first so I could scoop up
- 1:02:53Open AI #1, Inflection #2, Character A, Character AI #3,
- 1:02:59Glean #4, and Mistral AI #5. Max you want to read yours?
- 1:03:05Yeah. My team is Databricks, Pine
- 1:03:07Cone, Cohere Modular, and IMBU. Also known as the best team.
- 1:03:13And James. I went with Hugging Face,
- 1:03:16Anthropic, AI, 21 Labs, Replet, and Adept.
- 1:03:22That's our episode, and that's the Cerebral Valley AI series.
- 1:03:28Watch for YouTube videos. I'm sure I'll have Max and James
- 1:03:31once I'm like not fatigued of talking back to sort of reflect
- 1:03:35on it, but I'm not promising it immediately.
- 1:03:38But this has been a great. Run Thank you for sticking with
- 1:03:42us. And you missed some of the
- 1:03:44earlier episodes. Go back and listen to them.
- 1:03:48Thank you so much to Scott Brody, who's been the producer
- 1:03:51through all this and help help us figure out the show.
- 1:03:54So shout out to him. Thanks to Riley Kinsella, Gabby
- 1:03:57Caliendo, both key and sort of getting the conference and
- 1:04:02everything together. Thank you for everyone who's
- 1:04:05going to speak. And yeah, stick around.
- 1:04:10Please, like, comment, subscribe on YouTube, subscribe to
- 1:04:14newcomer.co and go play song quiz on your Alexa.
- 1:04:19All right? Yeah.
- 1:04:20Thanks so much. Thanks guys.
- 1:04:23Thank you all. Right.
- 1:04:24Bye.