The AI Startup Fantasy Draft

[00:00:04] Welcome to the AI Draft on the Cerebral Valley Podcast. I'm Eric Newcomer. We've got probably my favorite episode of the year coming up for you. We're about to pick what we think will be the most valuable AI companies in four years from now. This is a follow-up from our draft a year ago, but we'll catch you up on everything that happened last year. We'll be picking a bunch of new companies. I'm here with James Wilsterman.

[00:00:30] Hey, Eric. Happy to be here. Excited for the draft.

[00:00:34] James is the brain behind the draft. He'll explain it in a second. And Max Child.

[00:00:39] Hello, hello.

[00:00:42] Super excited. A very contentious game of forecasting the top companies in AI ahead of the Cerebral Valley AI Summit on November 20th in San Francisco. This is our way of thinking about all the top companies in AI before we host the AI Insider Event of the year.

[00:01:03] All that said, James, do you want to tell everybody the rules so they know why and how we're operating here?

[00:01:11] Buckle in. It's a complicated set of rules, so I will try to go through them efficiently. If you recall from last year, some of you may have listened to that episode. We drafted teams. As Eric said, we drafted the top private companies in AI.

[00:01:30] We're all looking to create a team of companies that by the end of 2028 will have the highest value on either sort of a private market valuation, an exit, or an IPO.

[00:01:43] So just to caveat a few different rules there, last year, we each drafted five teams. And in order to get the first pick, we agreed that you'd have to take a big handicap because there was such an outlier pick with OpenAI last year.

[00:01:59] So we agreed that Eric, by drafting OpenAI first, would have to take a $75 billion handicap on him.

[00:02:07] We had an auction for that. So I paid the most for the right.

[00:02:10] How are you feeling about that decision in retrospect?

[00:02:13] Excellent. Genius. I'm dominating right now.

[00:02:16] Yeah, Eric's team's looking really good. We'll get into that in a minute. A couple other rules.

[00:02:21] You have to be on this list at all to be one of the companies that you can draft. You have to have raised over $100 million.

[00:02:29] Fortunately, a lot of AI companies have raised well beyond $100 million.

[00:02:34] It's a long list.

[00:02:34] In a very short amount of time.

[00:02:35] Many companies we have never heard of.

[00:02:38] It's quite a long list. Many of them I haven't heard from since, but they have raised over $100 million.

[00:02:47] A couple other rules. You have to have had a core use of generative AI or other generative AI infrastructure use cases within your product.

[00:02:57] However, we're excluding certain industries as well.

[00:03:00] So if you're purely a biohealthcare industry, AI provider, a defense provider, silicon chips infrastructure provider,

[00:03:09] such as CoreWeave or Lambda Labs, we're also excluding you from this draft.

[00:03:13] Robotics and AVs as well. We're not drafting Waymo. We're not drafting Cruise and other AV companies.

[00:03:19] And we're also excluding a couple companies that are just purely based in China.

[00:03:24] We don't seem to know much about them and that makes it hard to draft them.

[00:03:28] And then, yeah, you have to have been founded within the last 10 years.

[00:03:33] So before, you know, excluding a couple companies that maybe have a generative AI story, but they were founded before 2013 or 2014.

[00:03:41] And then lastly, this year, there will be a way to add and drop companies.

[00:03:46] So you can actually decide in advance of this podcast recording whether you want to drop one of your companies off of your team.

[00:03:53] You can pick up new companies off of waivers.

[00:03:56] If you've ever played fantasy football, you'll be familiar with that concept.

[00:04:00] And then we also will extend the teams a little bit this year.

[00:04:04] So we're keeping our teams from last year, but we're adding two slots each to the team.

[00:04:09] So you'll have an opportunity to at least draft two more companies for your team.

[00:04:13] How's that sound, guys?

[00:04:14] Great.

[00:04:15] Great.

[00:04:15] I just want to emphasize, and I think you said this, but like the goal here is to get the highest total sort of market cap on your team.

[00:04:22] It's not like a return on your investment.

[00:04:24] So there's definitely an incentive to buy highly marked companies if you think, you know, there's still room to grow.

[00:04:31] But even if there isn't, you know, it may be strategic to pick a maximally valued company.

[00:04:37] The other thing I want to say is play along at home.

[00:04:39] Comment on the post with your picks.

[00:04:42] Obviously, you have the advantage of not competing with anyone else, but there are still so many companies that it's a challenge to come up with your...

[00:04:50] What's going to be the total number of companies we each have by the end of this, James?

[00:04:54] Seven.

[00:04:55] Yeah, seven each by the end of this waiver pick-up.

[00:04:57] Yeah, so pick your seven and let us know which companies you draft.

[00:05:02] Or let us know, you know, what were our genius picks and what were our terrible picks.

[00:05:08] Who are you rooting for, also?

[00:05:09] All genius.

[00:05:10] Team Dan?

[00:05:14] Newcomer subscribers are probably rooting for newcomer.

[00:05:16] Are they?

[00:05:17] I don't know.

[00:05:19] Maybe they're all rooting against newcomer.

[00:05:21] You don't know.

[00:05:21] Could be a hate watch.

[00:05:24] Revenge for scooping their fundraiser.

[00:05:26] What else do we need to cover?

[00:05:27] Anything else?

[00:05:28] All right.

[00:05:29] So a lot of rules, but we know there are a lot of nerds on here who want to make sure, understand the strategy behind it and follow along very closely.

[00:05:39] So those are the rules.

[00:05:40] And now we get to the fun part and make our picks.

[00:05:43] Give it a listen.

[00:05:44] This episode is presented by Brex, the financial stack that founders and VCs can truly bank on.

[00:05:51] Imagine what your founders could do with their runway if they had a banking solution that had no minimums, no transaction fees, and 20 times the standard FDIC protection.

[00:06:00] Plus, they could earn an industry-leading yield while maintaining access to funds whenever needed.

[00:06:07] Brex simplifies financial services for startups so they can focus on building.

[00:06:12] Connect your portfolio to the financial stack that one in three U.S. venture-backed startups already use.

[00:06:18] Check out brex.com forward slash banking dash solutions.

[00:06:25] The draft started out strong with a big pick from Eric, OpenAI, but he had to take a handicap on that.

[00:06:33] So he had to sacrifice seven-

[00:06:36] I insisted on paying to guarantee that I would have the right to pick OpenAI, and I've been extremely validated by that already.

[00:06:43] Yeah, so you took a $75 billion handicap off of your own total score.

[00:06:48] And by the way, this draft ends in November 2028.

[00:06:53] However, you, Eric, have also had some exits or semi-exits, both inflection and character AI semi-exited.

[00:07:01] We've decided to award you their valuations at the time of exit.

[00:07:06] Do you want to talk about that nuance for the listeners?

[00:07:09] A key subplot of this year has been sort of the licensing deals, acquihire, somewhat exit.

[00:07:16] Basically, leadership at a company gets bought.

[00:07:20] Shareholder, you know, the acquirer, sort of acquirer, pays the company, the startup, a big licensing fee.

[00:07:27] Then the startup uses that money to buy out the preferred shareholders, basically wiping the cap table.

[00:07:34] So it's a weird- The company still exists, but the investors got paid out and the leadership's gone.

[00:07:40] And so, you know, I think the logic we came to agree on was that these were effectively exits for the investors,

[00:07:47] and therefore you should be able to cash out.

[00:07:49] So the licensing deals, character, which I think we're putting at $2.5 billion, inflection for me.

[00:08:00] Inflection around $1.4 billion.

[00:08:01] And then, what, James, you had a debt?

[00:08:03] Yeah, at around a billion.

[00:08:05] Yeah.

[00:08:06] And I guess in the abstract, you could rebuy those companies,

[00:08:10] though we've also put in the requirement that companies need to have raised $100 million.

[00:08:14] Are we allowing rebuys of these?

[00:08:16] I'm never going to rebuy any of them, to be clear.

[00:08:18] I don't think we should allow rebuys because, like you said to us,

[00:08:23] I believe reported that the cap tables were kind of completely wiped out.

[00:08:28] So these companies haven't-

[00:08:29] The new shells of the former companies haven't really raised $100 million yet,

[00:08:33] so they're no longer part of the draft.

[00:08:36] Your other companies are doing quite well.

[00:08:38] So yeah, OpenAI reportedly raised $157 billion valuation.

[00:08:43] Glean raised an E round at a $4.5 billion valuation.

[00:08:49] Mistral also raised a B round around $6 billion.

[00:08:54] So your total, including the $75 billion handicap, is roughly around $100 billion right now, 96.4.

[00:09:03] So you are way ahead of Max and I.

[00:09:06] Max had Databricks, Pinecone, Cohere, Modular, and Embue.

[00:09:12] I'm using to value Databricks because it hasn't raised a new round since, but Databricks is an extremely valuable company.

[00:09:20] So Max is worth at least half of what I am.

[00:09:23] There's no way.

[00:09:24] So thank you.

[00:09:26] Thank you, Eric.

[00:09:27] I appreciate the defense of my honor here because the spreadsheet says my total value is five compared to your 96.

[00:09:34] So yeah, it's like I was like, is this just going to be a dunk on Max Fest?

[00:09:40] Just because Databricks still looks pretty good to me.

[00:09:42] Yeah, definitely.

[00:09:44] You need to get a new fundraise to get points on the board, but I agree.

[00:09:49] You're still holding strong.

[00:09:50] But your only points on the board would be Cohere, which raised a Series D and came in at about $5.5 billion.

[00:10:00] And then my own team, I have Hugging Face, Anthropic, AI21 Labs, Replit, and Adept.

[00:10:07] As we said, Adept kind of exited at that $1 billion mark.

[00:10:10] And then Anthropic also did a fundraise in the last year that valued it at an $18.4 billion post-money valuation,

[00:10:19] putting my current score around $19.4 billion.

[00:10:24] And Anthropic is reportedly up to raise it from $30 to $40, so you'll get a boost.

[00:10:29] I was going to say, James and I are kind of in the same category of like we're around $40 probably

[00:10:35] because I think Databricks is probably around $40 and Anthropic is probably around $40.

[00:10:39] Yeah.

[00:10:39] And Eric is crushing us at $90 for sure.

[00:10:42] Yeah, sure.

[00:10:43] My roster is pretty weak.

[00:10:45] I went with the stars in Scrubs strategy, except I only have one star,

[00:10:48] and that is Databricks, something like $40, $40, $50.

[00:10:51] I have Pinecone, Cohere, which hung in there at a $5.5, as James said.

[00:10:56] I have Modular, and I have Imbue.

[00:10:59] The Scrubs are still scrubbing pretty hard, so there may be some foreshadowing about it.

[00:11:04] You biased towards some people you interviewed on stage, which was perhaps a mistake.

[00:11:10] I did go with a personal connection vibes-based approach.

[00:11:15] It may not have panned out, but I got Databricks, so I'm arguably still in second place, and I will take that.

[00:11:22] All right.

[00:11:22] Let's go into the drops.

[00:11:25] So, Eric, you freed up two spaces already with your inflection and character sales.

[00:11:30] Are you dropping any other companies?

[00:11:33] I'm not.

[00:11:34] I'm not dropping anything else.

[00:11:35] I like, obviously, OpenAI, clean, mistraw.

[00:11:39] I'm keeping.

[00:11:40] I want to say one thing.

[00:11:41] You know, I was very bullish, I think, professionally on OpenAI.

[00:11:45] I just wrote a bare column on OpenAI at $157.

[00:11:50] Max and James are not giving me a way to somehow sell my current markup.

[00:11:56] I would never drop OpenAI.

[00:11:58] You can exit or IPO.

[00:12:00] But I'm not allowed to lock in where it is and then sort of be short.

[00:12:05] I'm still going to go up.

[00:12:07] I think it's going to get marked up again.

[00:12:09] I think it'll get marked up.

[00:12:10] Got it.

[00:12:10] But you're just a little.

[00:12:11] $40, $50 billion, but I'm skeptical beyond that.

[00:12:15] Got it.

[00:12:16] I would say if with OpenAI, this may be a future debate, but every new round is getting increasingly bizarre in terms of structure.

[00:12:23] Right, exactly.

[00:12:24] Headline valuations, yeah.

[00:12:26] I talked to some investors in OpenAI round and they're like, well, we get paid back in two years if they don't grow fast enough.

[00:12:32] Exactly.

[00:12:32] It's starting to become like a sort of debt-like structured security.

[00:12:36] I know.

[00:12:36] If I could cash out for the money it's worth, I would be tempted.

[00:12:40] But I want to be able – given I'm keeping it, I'm keeping it with pride and I'm going to pat myself on the back if I do it all with it.

[00:12:46] All right.

[00:12:47] Well, if the next OpenAI round involves collateralized obligations or something, we need to start talking about what a valuation even is at this point.

[00:12:56] Right.

[00:12:57] So, whatever.

[00:12:58] There's a lot of nuance in this VC game, I think is what we've discovered from this draft.

[00:13:03] Who knew?

[00:13:04] Who knew?

[00:13:05] I thought it was easy.

[00:13:06] Yeah.

[00:13:07] Okay.

[00:13:08] Max, are you dropping anyone from your team?

[00:13:11] I am dropping some of my scrubs.

[00:13:14] Sad.

[00:13:15] I will – yeah.

[00:13:16] RIP Max's team.

[00:13:18] I will – I was honestly considering dropping everyone except Databricks, but I decided to dial it back from that approach.

[00:13:27] I will stick with Cohere given I'm starting to believe that –

[00:13:32] Yeah, the markup.

[00:13:33] Also, I think foundational models are becoming state-level assets and potentially Canada will want their own foundation model.

[00:13:40] You know, like Safe Superintelligence might be Israel's foundation model.

[00:13:45] So, I'll keep Cohere and I will keep Modular because I still ride with my boy, Chris Latner.

[00:13:52] You know, it might not work out, but he's trying to create a new language.

[00:13:56] Yeah, okay.

[00:13:57] Yeah.

[00:13:57] So, I will be dropping Pinecone and Imbue.

[00:14:01] Pinecone I went kind of back and forth on, but the whispers in the VC circles I roll in quite regularly were not very positive.

[00:14:10] Imbue stole Volley's office, so there's a little bit of a bad blood.

[00:14:13] Imbue also stole Volley's office.

[00:14:16] Which you think is a positive sign.

[00:14:17] They've got a lot of money to throw around.

[00:14:20] Yeah.

[00:14:21] Outside of the bad blood on the office front, Eric, would you agree?

[00:14:24] Imbue, not a lot going on since I picked them.

[00:14:27] I haven't heard much.

[00:14:27] I haven't really heard much out there.

[00:14:28] They're not on our radar as much.

[00:14:29] Yeah.

[00:14:30] So, either they're working on the greatest takeover of the world strategy ever, you know, or there's not a lot going on there.

[00:14:36] So, I'll go with not a lot going on.

[00:14:37] I'm dropping those two.

[00:14:38] Pinecone, Imbue.

[00:14:39] I wish I could say it was fun, but it wasn't really.

[00:14:42] All right.

[00:14:42] All right.

[00:14:43] I'm going to put an upward.

[00:14:45] My team, I'm feeling good about Anthropic.

[00:14:48] Anthropic, as we've said, I think, you know, even in the last year, we've seen a lot of kind of progress in terms of catching up to OpenAI in the closed model race.

[00:14:55] And a lot of success with their, specifically their coding capabilities and enterprise adoption.

[00:15:03] So, very excited about Anthropic.

[00:15:06] Hugging face as well.

[00:15:07] Still, I would say core player in the space.

[00:15:09] So, I'm keeping them.

[00:15:10] I was decided to drop AI21 Labs.

[00:15:14] They essentially have an open source model, but it seems pretty behind other open source providers.

[00:15:20] And they, you know, their claim to fame is a lot to do with context windows and kind of inference time.

[00:15:26] But I think that is getting a little bit commoditized, as we've discussed in earlier segments here.

[00:15:33] So, dropping them.

[00:15:35] I was debating Replit, but I've decided to keep them on my team.

[00:15:40] Oh, you're keeping Replit.

[00:15:41] Okay.

[00:15:41] That was interesting.

[00:15:42] Give us the case for why you're keeping Replit here.

[00:15:47] So, Replit is hot.

[00:15:50] Replit, if I were starting as an engineer again, I would probably be using them.

[00:15:55] I think that it is the best way to get started coding.

[00:15:59] They have a vision of, you know, empowering like billions of, you know, creators to create applications.

[00:16:06] I think it's a little bit of like a long run or a high upside bet.

[00:16:11] And I still think it's the right – they have, you know, product market fit.

[00:16:17] They have customers.

[00:16:17] They have 20 million users or something.

[00:16:19] They're working on AI agents.

[00:16:21] They're really in the mix.

[00:16:22] I hear about them a lot, even on X.

[00:16:25] So, I think there's – yeah, there's still some upside to be had there.

[00:16:30] They're good at – it was a tough call.

[00:16:32] They're out there.

[00:16:33] Prominent on X.

[00:16:34] Yes.

[00:16:34] I was – can I –

[00:16:35] What do you think?

[00:16:35] Is that a better take?

[00:16:37] My contrarian take here – well, not contrarian at all, but contrarian to you apparently – is I think what we learned about the first year is we are not in the unicorn hunting business.

[00:16:47] We're in the decacorn hunting business.

[00:16:49] Yes.

[00:16:50] Because nothing shows up on this chart other than the $10 billion plus valuations.

[00:16:56] Right.

[00:16:56] You guys, congratulations.

[00:16:58] You had exits.

[00:16:58] Like real VCs.

[00:16:58] Like real VCs.

[00:16:58] They're both like $1, 2, 3, 4 billion, whatever.

[00:17:01] Yeah.

[00:17:01] Exactly.

[00:17:02] Well, I think real VCs, if you're a seed VC, you can be in the unicorn hunting business.

[00:17:05] Sure.

[00:17:05] But we're basically gross stage VCs because we're $100 million minimum.

[00:17:10] So, we're in the decacorn hunting business.

[00:17:12] $3 billion, $10 billion.

[00:17:13] Yeah.

[00:17:13] I just – yeah.

[00:17:16] Right.

[00:17:17] So, I'm just not sure.

[00:17:18] I'm not sure I see the path to decacorn for replet.

[00:17:22] How would you see that for modular?

[00:17:25] Explain the decacorn method for modular.

[00:17:27] Okay.

[00:17:28] Here's the decacorn case for modular.

[00:17:30] Modular replaces CUDA as the core language for how you code and gets bought by like AMD or ARM or something as like the way you code on non-NVIDIA chips.

[00:17:41] And there's so much demand for getting off NVIDIA, not paying Jensen, exorbitant fees that there's huge value in having an alternate AI coding ecosystem.

[00:17:50] So, there you go.

[00:17:51] That's what I believe.

[00:17:52] Look, I think it's a very asymmetric upside bet where it would probably be worth nothing in 90% of cases.

[00:17:58] But I like my 10% upside case.

[00:18:00] And I'd be interested in your 10% upside case for replet.

[00:18:02] Like give me the decacorn replet.

[00:18:03] Well, no.

[00:18:04] It's basically that that becomes how everyone learns to code.

[00:18:09] The number of engineers in the world.

[00:18:12] Seven.

[00:18:13] I mean, there's no – everyone codes in the world.

[00:18:16] Like the tan of that is unclear.

[00:18:17] Sure.

[00:18:18] You didn't let me finish.

[00:18:19] I think that let's say you have essentially hundreds of millions or billions of engineers in the world.

[00:18:28] Like it becomes the spreadsheet or something, right?

[00:18:32] Coding.

[00:18:32] Everyone became like sort of an engineer in the PC era through spreadsheets.

[00:18:40] But maybe true application building becomes like a core part of people's knowledge workers' jobs in the future.

[00:18:46] And, yeah, maybe it becomes a true platform for that.

[00:18:48] Or there's a bit more of like a consumer take, I guess, where somehow creating these applications becomes part of a social network type experience.

[00:18:57] But, yeah, I think it's a stretch.

[00:18:59] But it's possible, I think.

[00:19:01] It looks like we all have four spots to fill given our drops and our new slots.

[00:19:07] So we're going to pick up exactly where we left off, the snake draft.

[00:19:11] I had the last pick of the draft last year with Adept at number 15.

[00:19:16] So it'll be my pick coming in hot with draft pick number 16.

[00:19:22] And then we'll snake draft from there to Max and Eric.

[00:19:26] So, yeah, my pick.

[00:19:28] I don't think this will come as much of a surprise, but I'm picking XAI, which has been rumored to be raising at a $40 billion valuation.

[00:19:39] Did you give it an extra premium for being anti-woke AI?

[00:19:43] Well, Eric really sold me that they were going to take over the world.

[00:19:47] So very high.

[00:19:49] Yeah, Eric really boosted.

[00:19:51] Yeah.

[00:19:52] No, it's a good pick.

[00:19:53] Defend your obvious pick.

[00:19:54] Yeah, go ahead.

[00:19:55] I think it's a good pick.

[00:19:57] I think they spun up, I think, 100,000 H100s in the last four months, which Jensen himself said wasn't possible.

[00:20:05] But then, according to Elon, they've done it.

[00:20:08] They appear to be trucking in gas-powered generators to power this thing.

[00:20:17] They're just going as fast as possible, I think, which is pretty compelling.

[00:20:21] And kind of what you need to do to be a foundation model competitor in the environment today.

[00:20:25] So, yeah.

[00:20:27] Don't short Elon, I guess.

[00:20:29] And, you know, they have that.

[00:20:31] Unless it's Twitter, in which case you probably should have shorted Elon.

[00:20:33] But, yeah.

[00:20:34] Yeah.

[00:20:35] That's true.

[00:20:36] AI does feel like more of a hardware problem and an energy problem than, like, a social networking problem.

[00:20:42] More in his wheelhouse.

[00:20:43] Yeah, right.

[00:20:44] And maybe XAI helps Twitter, right?

[00:20:47] Yeah.

[00:20:48] Maybe.

[00:20:48] I don't know.

[00:20:49] Yeah.

[00:20:49] I think, or vice versa.

[00:20:51] I think it's kind of an interesting...

[00:20:53] I'm actually very intrigued by how that works from a business perspective.

[00:20:57] Because the current only product of XAI is this Grok chatbot that's available only in X Twitter.

[00:21:04] So, I don't know how this looks as a partnership long-term.

[00:21:08] Maybe they become the same company or something.

[00:21:10] But this feels like a good pick to me.

[00:21:14] I think it's the obvious first pick.

[00:21:16] I think the short case is it's, like, the third best model right now.

[00:21:19] But the long case is Elon's good at bending the laws of physics.

[00:21:23] And AI is mostly a laws of physics problem in some ways with, yeah, chips and energy.

[00:21:26] And recruiting.

[00:21:27] And raising, fundraising.

[00:21:29] Yeah.

[00:21:30] Great pick.

[00:21:30] Totally.

[00:21:31] Yeah.

[00:21:32] All right.

[00:21:33] All right, Max, you're up next with the next pick of the draft.

[00:21:36] I will take kind of the obvious second pick.

[00:21:39] I set up Eric whining about how his OpenAI first pick screwed him for this redraft.

[00:21:46] I'll take Scale AI, which is far and away the second most valuable open slot on the board.

[00:21:52] They just raised $14 billion.

[00:21:54] They provide data labeling services to a bunch of different AI companies.

[00:21:59] I think what's interesting about Scale is they are very polarizing in Silicon Valley.

[00:22:04] I would say a lot of people have joked for many years that they're just a, like, sort of random third world worker, like, mechanical Turk business where they offload a bunch of menial tasks to people in cheap countries.

[00:22:23] But, on the other hand, they basically, like, 4x revenue in the last year.

[00:22:27] Some crazy thing from, like, 1 to 4 billion.

[00:22:29] And so, if I'm getting at this at, like, 3x revenue, I'm feeling pretty good.

[00:22:35] And, yeah, I mean, the lesson of the last draft is you've got to ride the horses that are winning.

[00:22:39] And this is a clear number two right now.

[00:22:41] So, I'll grab Scale.

[00:22:45] Somewhat conflicted about whether or not it's a long-term multi-deca-corn.

[00:22:49] But, you know, you can't not take the only other 10 billion plus valuation.

[00:22:54] Or the data piece of the AI puzzle, which we talked the least about but could end up being super cool.

[00:22:59] Yes.

[00:22:59] Yeah, we haven't really weighed in on the take that data, you know, quality of data is long run a huge piece of the puzzle, maybe equally or, you know, as important as chips and compute and scale and energy and so on.

[00:23:15] Because there's a lot of garbage data out there.

[00:23:17] And especially if you just scrape the whole internet and you don't have good stuff, you maybe won't get a good AI or same with, you know, imagery or video or whatever.

[00:23:23] So, that's the scale case.

[00:23:26] The scale short case, as I said, is that oddly enough, its business might be very automatable by AI in some ways.

[00:23:32] So, yeah, that's my take.

[00:23:33] And we have the CEO joining us at the conference.

[00:23:37] Yes, we have the CEO.

[00:23:37] And Eric, I'm sure we'll ask him some hard questions.

[00:23:39] So, it'll be fun.

[00:23:40] I'm excited.

[00:23:41] It'll be fun.

[00:23:43] I'm up.

[00:23:44] Eric, you're up.

[00:23:45] Yes.

[00:23:45] You get two picks.

[00:23:46] Oh, two picks.

[00:23:48] Okay.

[00:23:48] I think the first one is what you guys would pick.

[00:23:51] Partially hidden on the stock because its valuation increase, I think, is only reported and not locked in.

[00:23:58] But I'm picking perplexity.

[00:24:02] Smart.

[00:24:03] Yeah.

[00:24:04] I mean...

[00:24:05] Bad pick.

[00:24:07] Highest chance of disrupting Google search.

[00:24:11] I do think a lot of investors are very excited about it.

[00:24:15] Even though, you know, as a reporter, it gets me a little annoyed with their seeming story ripoffs, I do like the attribution philosophy and excited to see them continue to develop.

[00:24:27] I did not realize I was going to get two back-to-back.

[00:24:29] Can I get a genuine perplexity shortcase?

[00:24:32] Can you give one?

[00:24:33] Yeah, can I give one?

[00:24:34] Yeah.

[00:24:35] Can I actually critique this pick?

[00:24:37] While Eric picks his next pick.

[00:24:39] Go for it.

[00:24:40] I just think if you look at the data, like if you look at AppAni or you look at any of the open data services that tell you how much usage perplexity has, it's like shockingly small.

[00:24:51] It's like really, really not that popular.

[00:24:54] It's very popular among nerds on Twitter, which is a perfect overlap with venture capital.

[00:25:00] I know there's this hot take that it's going to disrupt Google or whatever, but most people are pretty just used to typing in the search box that's on their browser that's either on their iPhone or on their computer.

[00:25:12] And that's going to be Google forever and ever because they have the money to buy that box from Apple.

[00:25:18] According to the information, by the way, they have 20 million ARR.

[00:25:21] Is that surprising based on what you have seen?

[00:25:23] No, I mean, I think like, you know, they probably have a couple million MAU, you know, somewhere between two and five and you're monetizing those people, you know, reasonably effectively, but not crazily effectively.

[00:25:32] So that's ballpark.

[00:25:33] You know, yeah.

[00:25:33] So let's say they have five million MAU or whatever.

[00:25:36] Google has like two billion MAU or something or I don't know.

[00:25:39] It's it's you know, it's a thousand times bigger or five.

[00:25:42] And it's free.

[00:25:42] I know.

[00:25:43] And it's free.

[00:25:44] Yeah.

[00:25:44] So I just I just would say, like, I'm not convinced the differential against Google's AI summaries or whatever is going to stay there for very long.

[00:25:53] A bunch of people are going to rip off their interface innovations and I don't really see how they're going to get a distribution advantage.

[00:26:00] So, yeah, I'm definitely short at whatever the what's the latest valuation, like 10 or something.

[00:26:05] It's reported.

[00:26:06] Yeah, I'm sure I'm sure I'm sure I'm sure I'm probably long at like one or I don't know, but I'm I'm sure it next Google.

[00:26:14] Google.

[00:26:14] We'll see.

[00:26:15] No, I just think that I think it's an overrated take.

[00:26:17] Yeah, I'm sorry.

[00:26:20] Eric, who are you picking?

[00:26:22] You know, I went all over the place.

[00:26:23] I certainly have companies down lower on the valuation list that I'm probably more amped about.

[00:26:31] I thought about Sierra, like customer support being a sort of high, you know, a promising area where I could actually get deployed.

[00:26:41] But just looking at my strategy, like we're going for big ones.

[00:26:44] You got to go with the guy who is the genius behind OpenAI that could either lead research at another company, could invent AGI.

[00:26:53] If he invents AGI, I'm going to win.

[00:26:55] Like, so I'm going safe, super intelligence.

[00:26:57] Ilya, I'm betting on you.

[00:26:59] And then obviously, I mean, Daniel Gross is, you know, one of the most connected people in Silicon Valley these days.

[00:27:04] He's a co-founder.

[00:27:06] That's what I'm going with.

[00:27:07] Yeah.

[00:27:07] Yeah.

[00:27:07] For the record, I thought that was where you were going to say the obvious next pick was because that's definitely who I would have taken over.

[00:27:12] Interesting.

[00:27:13] Huh.

[00:27:14] Why would you get why are you guys so excited about this company?

[00:27:16] I'm short.

[00:27:17] Like, you're short.

[00:27:18] Right.

[00:27:18] I don't know.

[00:27:19] I guess as an owner of XAI.

[00:27:22] You believe in dangerous super intelligence.

[00:27:26] You're like, you're like, no, no, no.

[00:27:27] The super intelligence has to be dangerous to be valuable.

[00:27:30] Like, save is no good.

[00:27:32] I have a balanced, diversified strategy of having Anthropic and XAI.

[00:27:37] Exactly.

[00:27:39] Save the world, kill the world.

[00:27:40] Who cares?

[00:27:40] We can go safety.

[00:27:41] Give us the short case.

[00:27:43] Yeah.

[00:27:43] What's the short case?

[00:27:45] I just think that, yeah, like, I think that you're basically basing this off of, you know, a billion dollar valuation and he can create a new foundation model.

[00:27:58] Right.

[00:27:58] I think that's the only explanation for why this is important.

[00:28:01] And it just feels like he's so far behind.

[00:28:04] All right.

[00:28:04] So my pro safe super intelligence take here to stand behind Eric is I think that these foundation models are becoming so expensive that their borderline have to be nationalized in the next, like, two to four years.

[00:28:17] And I think safe super intelligence is basically going to be Israel's national model.

[00:28:23] At which point, you know, I don't know if that impacts the valuation and we can debate that a couple of years down the road.

[00:28:29] But fundamentally, like, it's going to be a nation state level concern to have a really high quality model.

[00:28:34] And I think Ilya will probably be the flag bearer of Israel, which maybe was the AI 21 case for James a year ago.

[00:28:42] But they might have lost out to SSI.

[00:28:44] So that would be my pro case.

[00:28:46] And that's why, yeah, I'm a fan of the pick.

[00:28:48] All right.

[00:28:49] Let's keep moving.

[00:28:50] All right.

[00:28:51] Max, you're up with the next pick.

[00:28:55] Well, Eric kind of foreshadowed this that I think, you know, a solid option for next pick is Sierra, which is ex-Salesforce CEO Brett Taylor started a AI kind of customer support AI company.

[00:29:10] You know, I find customer support super boring.

[00:29:12] There's like 250 of these fucking companies.

[00:29:14] I don't really see any obvious reason these guys are going to win because I haven't seen the actual product or compared them.

[00:29:20] But like any good VC, I'm just riding the heat of all the other hot VCs being in this company.

[00:29:26] And everyone I talk to in the biz is like, this is it, baby.

[00:29:30] Benjmark and Sequoia can't be wrong.

[00:29:31] Yeah, exactly.

[00:29:32] Benjmark and Sequoia can't be wrong, baby.

[00:29:34] It's never lost me money before.

[00:29:35] So I'm going to freely admit I know almost nothing about why this is supposed to be the best of these 200 different customer support companies and say, take me to the top, Brett Taylor.

[00:29:46] I love it.

[00:29:46] Basically, you're just going with the VC cred.

[00:29:50] Hell yeah.

[00:29:51] Okay.

[00:29:52] I believe in that philosophy.

[00:29:55] That was my glean book last time, Kleiner and Sequoia.

[00:29:59] Totally worked.

[00:30:01] All right.

[00:30:03] I will take the next pick of the draft.

[00:30:05] I guess I get two picks here.

[00:30:08] I think this will be maybe a slight surprise.

[00:30:11] But with the next pick, I'm taking Eleven Labs.

[00:30:14] Ooh.

[00:30:16] Yeah.

[00:30:16] Eleven Labs.

[00:30:18] We are already a customer.

[00:30:19] They definitely have a lot of product market fit in this voice cloning and text-to-speech space.

[00:30:27] They seem to have been making moves, I think I alluded to earlier, buying companies that provide consumer applications so that you could read articles with these generated voices.

[00:30:41] They have released a lot of new technologies, including most recently, they announced kind of a new agent framework for voice AI applications.

[00:30:52] They also have a sound effects API.

[00:30:55] Yeah.

[00:30:55] They ship a lot of product.

[00:30:56] They have a lot of fans in the developer community.

[00:31:01] They're taking an eye towards some consumer applications.

[00:31:04] So I'm going with Eleven Labs.

[00:31:07] I'm short Eleven Labs.

[00:31:12] I just, as a customer, I like the product.

[00:31:16] It also seems immensely replicable by the large cloud services.

[00:31:19] I think once Lena Khan is no longer blocking acquisitions, it will get bought by AWS or Google or Microsoft.

[00:31:27] All right.

[00:31:29] And with my next pick of the draft, I'm going to take Runway.

[00:31:36] Man.

[00:31:37] All right.

[00:31:37] That was my next pick.

[00:31:38] Really?

[00:31:39] I would.

[00:31:40] I love them.

[00:31:41] Yeah.

[00:31:41] You're short.

[00:31:43] Fascinating.

[00:31:44] Okay.

[00:31:44] You think that's a bad pick, Eric?

[00:31:48] Yeah.

[00:31:48] I mean, he was the one in Cerebral Valley, New York, making promises that our own AI crowd didn't even believe.

[00:31:54] I don't know.

[00:31:54] And now they're sort of saying they're in like the video production business themselves.

[00:32:01] I'm very bullish on Adobe figuring out what creatives want and monetizing it well.

[00:32:06] Yeah.

[00:32:06] So, no, I'm not.

[00:32:08] I just think we're in the business of the Decacorn hunting runway.

[00:32:13] Anyway, the hype train is why I'm investing in them here.

[00:32:17] They said we'd be creating Oscar winning fully feature length movies in two years or something.

[00:32:24] They told Max that on stage.

[00:32:26] I think there is a lot of high upside scenarios where video gen tools become like a much more interactive medium also.

[00:32:36] Right?

[00:32:37] So, if you could really generate video that's photorealistic and sort of in that low latency, you can create almost gaming use cases.

[00:32:43] There's a lot of like simulation type products they could build.

[00:32:47] And then there's consumer use cases, just, you know, creator tools.

[00:32:50] So, they do seem to be the leader in this important AI video space.

[00:32:54] We also haven't seen, you know, OpenAI launch Sora, even though that received a lot of, you know, attention when they announced it.

[00:33:01] Maybe because it's super expensive.

[00:33:04] Yeah.

[00:33:05] Do you think runway is just burning, burning capital then?

[00:33:08] I don't know that specifically.

[00:33:09] I'm just saying if all these models cost a ton, like video models sound like they're going to be like, you know, so, so cost intensive.

[00:33:17] Yes.

[00:33:18] And right now, I think, you know, the main customers of runway are like feature length Hollywood, you know, production studios or maybe like doing, you know, a few like one second clip generations or something to like fill in part of their films.

[00:33:32] But I think that, you know, as cost curves come down, it just opens up a lot of use cases.

[00:33:36] But yeah, if cost doesn't come down, then maybe it's a bad pick.

[00:33:39] I like the asymmetric upside bed of take over the world plan.

[00:33:41] So, I'm kind of in.

[00:33:43] Should I take the next pick?

[00:33:44] Yeah.

[00:33:44] Yeah, go for it.

[00:33:45] All right.

[00:33:45] We're definitely getting into the harder part of the draft where I think you have to go a little more wild with a little crazier pick.

[00:33:52] I'm going to take Sakana AI from one of my VC referrals here said, this is the hot thing.

[00:34:01] You should check it out.

[00:34:02] Apparently, they're literally trying to use AI to automate fundamental scientific discoveries.

[00:34:09] Essentially, they're huge in Japan.

[00:34:11] I believe in basically the core mission that it is possible to make scientific breakthroughs with AI.

[00:34:18] And so, I don't know.

[00:34:19] That sounded like an asymmetric high upside bet.

[00:34:21] And I'm getting good buzz from the VC networks.

[00:34:23] And at this point, it's pretty hard to kind of dig through the draft.

[00:34:25] So, Sakana, welcome to the team.

[00:34:28] Is it okay if I show my hand and sort of talk through the four and then pick two?

[00:34:32] Yeah, sure.

[00:34:33] I mean, yeah.

[00:34:33] You're just going to leak ideas to us.

[00:34:36] So, it's all good.

[00:34:37] Yeah.

[00:34:37] Maybe I won't.

[00:34:38] No, no, no.

[00:34:39] Come on.

[00:34:40] Is it a podcast first or is it winning first?

[00:34:42] Do it for the content, baby.

[00:34:44] All right.

[00:34:45] I'm debating.

[00:34:45] Thesis one is like, man, we keep saying it's all about applications.

[00:34:50] We're bullish on coding.

[00:34:54] And the other big hyped application category is legal.

[00:34:58] So, Codium is coding.

[00:35:00] Harvey is legal.

[00:35:02] Those would be applications.

[00:35:04] Cursor is very buzzy, but I'm not allowed to take it.

[00:35:06] They haven't raised $100 million.

[00:35:08] I met an interesting services legal startup whose name I'm blanking on.

[00:35:12] Sorry, guy.

[00:35:13] But I don't think they've raised enough money.

[00:35:16] So, that would leave Harvey and Codium.

[00:35:18] Weights and Biases, I think, is a company that's done well in serving a lot of the infrastructure layer and could be a solid company.

[00:35:27] Listening to myself, I'm clearly excited about getting hits on coding and law.

[00:35:34] So, I'm going to take Codium.

[00:35:36] I'm going to take Codium first and then Harvey second.

[00:35:43] All right.

[00:35:44] All right.

[00:35:44] What do you think of the Harvey controversy that –

[00:35:47] What do you think of the Harvey controversy that Harvey actually sucks?

[00:35:49] Well, there's that random tweet that is like vaporware.

[00:35:52] You talk to some of the legal startups and they're like, we don't see them out there.

[00:35:56] Well, I've talked to a couple lawyers and they're like, it's not that great right now.

[00:36:01] Now, maybe you're like, I'm long models getting better, so it doesn't matter.

[00:36:04] They have a better sales force, yada, yada, yada.

[00:36:06] But –

[00:36:06] With both of them, I'm betting on the category without knowing a lot about the company.

[00:36:10] Sure.

[00:36:11] You're just basically trying to grab the leader in a category you believe in.

[00:36:14] All right.

[00:36:15] Exactly.

[00:36:15] Fair enough.

[00:36:16] Yeah.

[00:36:16] I think that's a reasonable thesis.

[00:36:18] I do think the buzz that Harvey is not that great is it seems to have some where there's smoke, there's fire kind of element to it from my research.

[00:36:27] What makes you so sure that Codium is the leader in this category?

[00:36:33] Who's my other option on this list?

[00:36:36] Well, I don't know.

[00:36:37] I'm just wondering what's the evidence of that.

[00:36:39] They just raised enough money to get on the list, I guess is what I'm saying.

[00:36:43] Isn't magic the same thing kind of or similar?

[00:36:47] Yeah.

[00:36:48] And obviously, Cursor hasn't raised enough, but magic was like Daniel Gross' play into that space, right?

[00:36:53] Yeah.

[00:36:54] I've heard of Codium.

[00:36:56] Fair enough.

[00:36:57] The end of the draft is tough, man.

[00:36:59] I think you're just grabbing ideas here.

[00:37:01] I mean, Cognition is also Codium, to be clear.

[00:37:04] Yeah.

[00:37:04] I could have taken that.

[00:37:06] Yeah.

[00:37:06] Sure.

[00:37:07] Exactly.

[00:37:07] Exactly.

[00:37:08] Cognition.

[00:37:08] Yeah.

[00:37:08] Poolside is Coding.

[00:37:10] Sourcegraph is somewhat code related.

[00:37:13] There's just a lot of coding companies up here, so I was just curious, but yeah.

[00:37:18] Yeah.

[00:37:19] Why you went with Codium specifically, but it's all good.

[00:37:22] It's all good.

[00:37:23] It's done.

[00:37:24] It's done is done.

[00:37:25] All right.

[00:37:25] I will.

[00:37:26] It's done.

[00:37:27] What's done is done.

[00:37:28] The bottom of the draft is tough.

[00:37:31] Max, your last pick.

[00:37:32] My final pick.

[00:37:35] I will take Hebbia, which is sort of like Eric's legal thesis, but for Excel in some

[00:37:44] ways.

[00:37:45] It's sort of like the finance, like private equity, investment banking to some degree,

[00:37:52] whatever.

[00:37:53] Excel jockey AI, basically.

[00:37:54] I've heard very good things about the product.

[00:37:57] It has a super elite investor list with Peter Thiel and Andreessen and everything.

[00:38:02] And I actually talked to someone who worked there and they were like, this is a pretty

[00:38:06] sick product.

[00:38:06] And I don't think they were just talking their own book.

[00:38:08] So yeah.

[00:38:09] Hebbia.

[00:38:10] AI for Excel jockeys, which I was an Excel jockey 12 years ago.

[00:38:14] So that sounds pretty good to me.

[00:38:15] Would never have considered that in my life.

[00:38:19] It's the bottom of the draft.

[00:38:21] It's the bottom of the draft.

[00:38:22] You got to take some sleepers.

[00:38:24] You got to have a dream.

[00:38:30] All right.

[00:38:30] Last pick of the draft, James.

[00:38:32] Last pick of the draft.

[00:38:33] A lot of companies still on the board.

[00:38:37] Yeah.

[00:38:37] So I'll walk you through some that I'm considering.

[00:38:42] Kind of a debate between, I guess, three companies here.

[00:38:46] Poolside, which I mentioned is another AI-powered software development platform.

[00:38:50] They raised $500 million in a Series B this month, which I thought was pretty interesting.

[00:38:56] Staggering quantity of money for other purposes.

[00:39:01] Theoretically for a foundation model trained around coding.

[00:39:07] The other one on my list still, I think Eric mentioned them earlier.

[00:39:10] They do seem to be kind of a leader in this developer space where if you wanted to use

[00:39:18] open source models and do fine tuning and host them and have the infrastructure for that,

[00:39:23] they seem to be one of the leaders in that space.

[00:39:25] And then the last one would be very much a flyer, which is just World Labs, which is

[00:39:31] Fei-Fei Li's startup, which just raised about $100 million in July at a $5 billion valuation.

[00:39:38] So, yeah, there's some opportunity there.

[00:39:43] But I think similar to my bare case on safe superintelligence, I think I will be rolling that one out.

[00:39:52] And in terms of going after that sort of software space, I think I'm going to have to pick Poolside.

[00:40:01] Oh, wow.

[00:40:02] Lots of money.

[00:40:04] They could sell for a billion, you know, just based on the money.

[00:40:08] Just for the cash.

[00:40:08] Just to get the cash.

[00:40:09] You know, I mean, yeah, I don't know.

[00:40:13] I don't know what they're working on.

[00:40:15] I hope it's worth a lot.

[00:40:18] High upside.

[00:40:19] High upside.

[00:40:20] High upside.

[00:40:21] Yeah.

[00:40:22] Should we do some Bring Out Your Dead?

[00:40:23] Any picks that we...

[00:40:24] We think you should have picked?

[00:40:26] You should have picked?

[00:40:27] Yeah.

[00:40:27] I mean, I don't know.

[00:40:28] I'm shocked you didn't pick Suno, James.

[00:40:29] You're just banging on Suno like every weekend.

[00:40:32] Suno's a great product.

[00:40:33] I thought you were going to grab some Suno stock.

[00:40:36] I just...

[00:40:37] I'm very nervous about the music industry, having worked with the music industry.

[00:40:40] You worked with the music industry.

[00:40:42] You have some questions about...

[00:40:43] Yeah, when do we get to start picking Volley in this?

[00:40:45] Like, oh, no, you haven't raised...

[00:40:47] Next round.

[00:40:48] We haven't raised...

[00:40:48] Next round.

[00:40:49] Yeah.

[00:40:49] Yeah, that'd be good.

[00:40:50] Yeah.

[00:40:51] I think Magic is interesting.

[00:40:53] I like that as another coding pick.

[00:40:56] My VC Whisper Source hated Together.

[00:40:59] Hated Together.

[00:41:00] Oh, yeah.

[00:41:00] That was interesting.

[00:41:01] I didn't go in Together then.

[00:41:03] Yeah.

[00:41:03] He was like, I think Together's a zero.

[00:41:04] Max, you're behind, so you're doing more research now.

[00:41:09] No, yeah.

[00:41:09] I feel like last time I did a lot of research, and this time I was just like...

[00:41:13] You're like, I'm winning.

[00:41:14] I'm coasting.

[00:41:15] I'm winning, you know?

[00:41:17] I'm not sure about these also-ran kind of models we have here.

[00:41:20] I think Pika is the third best video model, so I'm a little concerned about that.

[00:41:27] Cognition's interesting with Devin.

[00:41:29] Cognition, yeah.

[00:41:30] Yeah.

[00:41:30] Yeah.

[00:41:30] But I guess it seems to be less of a foundation model play and more of this agentic play,

[00:41:37] but maybe the hype has died down a little bit.

[00:41:40] It is amazing how just in six months, the hype cycles are just changing so rapidly that

[00:41:45] stuff that was just crazy hot at our last conference or two conferences ago now feels

[00:41:49] like, oh, that's tired.

[00:41:51] They haven't done anything in eight months or something, and it's like, okay, well,

[00:41:55] the cycle time in the AI industry is just ludicrously short right now.

[00:42:02] Cool.

[00:42:03] Cool.

[00:42:04] Should we wrap up our picks?

[00:42:05] All right.

[00:42:06] Yeah.

[00:42:07] Go ahead.

[00:42:07] Yeah.

[00:42:09] Okay.

[00:42:10] Just let's compile this quickly.

[00:42:12] So, Eric, your team now consists of OpenAI.

[00:42:17] You banked some money from Inflection and Character, and you continued to ride with Glean and Mistral,

[00:42:25] and then you added four picks to your team.

[00:42:31] Perplexity, great pick, I think, but risky, as Max said.

[00:42:36] Safe superintelligence.

[00:42:38] Safe pick.

[00:42:40] Safe, safe, safe pick.

[00:42:43] Safe and superintelligent pick.

[00:42:44] Yeah, safe super pick.

[00:42:47] Kodium and Harvey.

[00:42:49] How are you feeling, Eric, about your team?

[00:42:51] I'm happy, honestly.

[00:42:52] I think, you know, I wish I could have gotten XAI, but given what was available, I'm happy with my picks.

[00:43:03] Max, you dropped a few folks from your team.

[00:43:07] You kept Databricks, Cohere, and Modular, and then you added ScaleAI, high valuation play there.

[00:43:17] You added, let's see, sorry.

[00:43:23] Sierra.

[00:43:24] Sierra.

[00:43:25] Sierra, Sakana, and Hebbia.

[00:43:26] Sakana and Hebbia, yes.

[00:43:28] I like things that end in A, apparently.

[00:43:32] And start with S.

[00:43:34] And start with S.

[00:43:35] Yes.

[00:43:39] And then, how are you feeling about your team?

[00:43:42] I mean, I'm feeling better than last year.

[00:43:44] I'm glad I banked Scale, so I've got some valuation gains.

[00:43:48] You still have Databricks, right?

[00:43:49] I still have Databricks, I still have Scale, I've got some valuation.

[00:43:52] You know, I still think I'm basically running the Stars and Scrubs play here, which didn't really work last time.

[00:43:59] So, other than, again, just Databricks.

[00:44:02] So, we'll see.

[00:44:04] We'll see.

[00:44:05] I'm a little less optimistic, perhaps, than I was last time around.

[00:44:09] And then, my team, I have Hugging Face, Anthropic, Replit.

[00:44:16] I banked some cash from the Adept Amazon deal.

[00:44:19] And I'm adding XAI with the highest valuation on the board right now.

[00:44:26] Adding Runway, adding Eleven Labs, and adding Poolside.

[00:44:34] A lot of, like, sort of foundation models, developer tools.

[00:44:38] I think with Anthropic and Poolside, getting that kind of coding play.

[00:44:44] But also, Anthropic has one of the best, I would say, coding LLMs right now with Claude.

[00:44:51] But I am very bullish on that.

[00:44:54] That is a valuable use case of this technology.

[00:44:58] So, riding the developer tooling ecosystem.

[00:45:03] You're running, like, the murderer's row of companies I think are overrated, basically.

[00:45:08] Brutal.

[00:45:10] Which probably means I'm wrong.

[00:45:12] And one of them will be, like, a $200 billion company.

[00:45:14] Yeah.

[00:45:15] You only did one, necessarily, right?

[00:45:17] You're just naming them off.

[00:45:19] You're like, Poolside, Replit, Eleven Labs.

[00:45:23] What was the other one?

[00:45:24] Hugging Face.

[00:45:26] Hugging Face.

[00:45:27] Anthropic is great, for the record.

[00:45:29] Anthropic is immaculate.

[00:45:31] But you've just got a lineup of, like, Twitter memes that I hate, basically.

[00:45:36] Just because you feel like they're overvalued?

[00:45:38] Well, yeah.

[00:45:39] I just, like, I don't...

[00:45:41] Yeah, I think overvalued or just, you know, raise too much money in the case of Poolside with no product.

[00:45:46] Or I think the product has a sort of lack of, you know, large market.

[00:45:50] Would either of you pick Codium or Harvey?

[00:45:53] Were they on your list anywhere?

[00:45:55] Harvey would not be on my potential.

[00:45:57] Harvey was on my list.

[00:45:59] Harvey on your hater list.

[00:46:01] Yeah.

[00:46:01] Yeah.

[00:46:01] Yeah.

[00:46:02] Harvey was probably on my hater list.

[00:46:03] Yeah.

[00:46:04] So, in some ways, you want the company to be a little polarizing.

[00:46:06] You want people to be thinking about, right?

[00:46:08] Exactly.

[00:46:09] Exactly.

[00:46:09] Yeah.

[00:46:10] Is it polarizing, though?

[00:46:11] I bet the max anti-portfolio is better than my actual portfolio.

[00:46:16] So, don't worry.

[00:46:18] It's like, I need the inverse of my hater list will probably do better than my actual team.

[00:46:23] I mean, at the end of the day, I'm still, like, really exposed to OpenAI more than anything.

[00:46:27] It's like, yeah.

[00:46:29] Yeah.

[00:46:30] Yeah.

[00:46:31] Maybe the whole industry is.

[00:46:33] Yeah.

[00:46:34] I mean, that's been the question throughout the events.

[00:46:37] Just, is it really all just about OpenAI?

[00:46:40] Still to be seen.

[00:46:41] Right.

[00:46:42] One update from our overrated, underrated episode.

[00:46:47] A minor correction on one of the revenue figures.

[00:46:50] Max?

[00:46:51] Yeah.

[00:46:52] I would say a major correction.

[00:46:54] A major correction.

[00:46:55] Which makes me.

[00:46:55] We got the source.

[00:46:57] We relayed what the source said accurately.

[00:46:59] All right.

[00:47:00] It was outdated.

[00:47:01] All right.

[00:47:02] To make this clear, the information special pro subscription told us that Anthropic was

[00:47:08] making around $100 million in revenue per year.

[00:47:11] Apparently, CNBC reported that Anthropic is, in fact, closing in on making a billion dollars

[00:47:17] in revenue a year, which is quite a bit larger, and is growing its revenue at basically 10x year

[00:47:25] on year.

[00:47:25] So, that $100 may have been accurate, you know, 9, 10, 11 months ago.

[00:47:30] Luckily for James and Eric, they both already look super smart, and they say Anthropic was

[00:47:35] underrated.

[00:47:35] We love it.

[00:47:36] It's much smaller revenue than it actually has.

[00:47:39] Yes.

[00:47:41] Max, luckily, gets to look super dumb when he said Anthropic was very overrated.

[00:47:45] Because I would say 10x year on year revenue growth absolutely justifies a crazy one third

[00:47:51] of OpenAI valuation.

[00:47:53] Interestingly, I stand by my take that their consumer app growth is not that impressive.

[00:47:58] I think it was 20% in the last quarter, because it looks like basically all the revenue comes

[00:48:03] from their API and their sort of developer tools.

[00:48:06] All right.

[00:48:07] So, that's a minor update on our overrated, underrated episode.

[00:48:11] Keeping ourselves honest, and we'll let you know if there's any new information that comes

[00:48:16] to light after these episodes.

[00:48:17] And yeah, reach out to us.

[00:48:19] You know, you can, what, email eric at newcomer.co probably.

[00:48:22] That's me personally.

[00:48:23] And send me your takes, comment.

[00:48:26] We want more comments.

[00:48:27] Yeah, comment on the post with your own opinions.

[00:48:29] And certainly happy to, we have very smart listeners.

[00:48:33] So, would love to hear where we got it wrong.

[00:48:36] Reach out anytime.

[00:48:37] Thanks so much.

[00:48:37] Reach out.