Latest / Newcomer Pod / Trump & AI Policy + Overrated/Underrated on the Cerebral Valley Podcast
Transcript
- 0:05Hey, it's Eric Newcomer. Welcome to the Cerebral Valley
- 0:08Podcast. I'm here with my good friends in
- 0:10Cerebral Valley, Co host Max Child.
- 0:13Hello. Hello, and James Wilsterman.
- 0:16What's up, Eric? Glad to be here.
- 0:18And we're going to talk about the miserable, happy, terrifying
- 0:22topic. The thing on everybody's mind,
- 0:24the election. We're going to try and focus a
- 0:27little on AI policy, but also just Silicon Valley's strong
- 0:31influence on the incoming Trump administration.
- 0:34James, how, how are you feeling right now?
- 0:37Well, I do feel like a bit of deja vu from 2016.
- 0:42And I would say I don't feel like shocked by the outcome in
- 0:45the same way, you know, I do feel like we're gonna we're just
- 0:48gonna go back on that roller coaster of Trump being so.
- 0:52We're resistance lips now. We're all sort of pink hats on
- 0:56or. I don't think we'll be it'll.
- 0:58I don't think the world will look like that, at least not in
- 1:01the immediate days of the Trump administration.
- 1:03But who knows what could happen, right?
- 1:06Max. Yeah, I mean, I think I had the
- 1:09same conversation like 8 times yesterday, which is everyone's
- 1:12like, I'm pretty upset. I'm also not surprised.
- 1:16I'm way less surprised than I was eight years ago.
- 1:20And to the resistance lid point, I think everyone's kind of like,
- 1:24you know, I'm tired of this crap and we're going to have to fight
- 1:28and slog our way through the next four years.
- 1:30But I, I also don't think it's going to go quite as you know,
- 1:33cuckoo pig hat direction as you were referencing.
- 1:36Well, I, I think there are a couple things, you know, and I
- 1:38wrote a big piece, but you know, one that Trump actually won the
- 1:43majority makes a little bit more and that the majority voted for
- 1:47him really knowing what he's about.
- 1:48So it's like, man, they this is what they wanted, you know?
- 1:52So I do think, I mean, none of us, you know, I think, yeah, I
- 1:55mean, I didn't vote for Trump. None of us voted for Trump.
- 1:57This is, you know, coastal elite crowd right here.
- 2:00But yeah, I mean, this is what America wanted.
- 2:04I mean, the, the point I would make is just that again, I mean,
- 2:07Kamal's going to probably run about 4-5 points behind Biden or
- 2:11something. That means one in 50 people
- 2:14switch sides, right? So if you imagine you're in a
- 2:16restaurant with 50 people, just one of them that voted for Biden
- 2:20last time switched over to Trump, right?
- 2:22Except in every single. Count except in every single
- 2:25state in every single district. So you could say, oh, the
- 2:26uniformity of that one in 50 switching is is surprising to
- 2:29me. Surprising.
- 2:30But but if you look around a restaurant and you're like,
- 2:32yeah, I can see one to two people in here switching size.
- 2:35You know, it's not that shocking.
- 2:36And I mean, I guess it's maybe surprising.
- 2:38That it happened everywhere, saw like a train wreck.
- 2:40That's what's surprising. It's like, get that, put that
- 2:43conductor back in there. But the data, the data shows
- 2:46that people think the country is a train wreck right now, right?
- 2:49We we all are like January 6th was a legendary train wreck.
- 2:52But the data shows more people think the country is a train
- 2:55wreck right now than like ever before.
- 2:58And whether or not you think that's right or wrong, that's
- 3:00how people are feeling. So one in fifty or two and 50
- 3:02switching is just, it's just not that surprising and I think
- 3:04everyone I talked to yesterday had that.
- 3:06And the prediction markets were right.
- 3:08I mean I I believe in the prediction markets, but I feel
- 3:10like the whale really made me skeptical.
- 3:13You were skeptical that that whale was driving a lot of that
- 3:17shift. Towards No, I believe, Yeah,
- 3:18yeah, yeah. I was skeptical on the results.
- 3:20Now I'm like, oh, he's smart. Well, it turned out there was a
- 3:22story that he had paid a pollster to run polls.
- 3:26You know, there's this strategy where you say, what does your
- 3:30neighbor believe, you know, and and that sort of reveals things
- 3:35that people otherwise might try to hide and that his sort of
- 3:38neighbor polling showed people even more supportive of Trump.
- 3:42And so it wasn't, you know, this guy was putting, you know, many
- 3:45millions of dollars to work on this bet.
- 3:47It makes sense that he he did his research, I mean.
- 3:51I'm not clear to me that his 30 million or whatever was actually
- 3:55driving the prediction market on Poly market.
- 3:57Like the liquidity is pretty high.
- 4:00I think a lot of people were betting on Trump.
- 4:01So I do wonder how much it was intelligent, you know, analysis
- 4:06versus like the crypto industry leans sort of towards
- 4:09Republicans. Anyway, there are endless,
- 4:11endless takes to be out on Trump.
- 4:13I'm sure you're going to, you know, the pod Bros or, I don't
- 4:17know, Matt Iglesias coming to us for, for our tech.
- 4:21So we will, we will now, now dig into what this means for AI
- 4:25policy. And tech policy in.
- 4:28General and tech policy. Yeah.
- 4:29I mean, I wanted to, like frame this up.
- 4:31I think, you know, we talked a little bit earlier about where
- 4:33we think the Trump administration can have a big
- 4:35impact on tech and AI specifically, and kind of laid
- 4:39out like 5 key areas that I think would be good to walk
- 4:41through. And obviously, if you guys feel
- 4:42like I'm missing anything, we'll throw it in there.
- 4:44But I think the big ones are energy policy.
- 4:47So, you know, nuclear, solar, how are we going to, you know,
- 4:51generate enough energy to power all these advancements in
- 4:54foundational models and and so on?
- 4:56I think chips is a huge one. CP, US, GP, US, NVIDIA, Taiwan,
- 5:02you know, TSMC in Arizona, all that kind of stuff.
- 5:05Regulation, as we discussed recently, the California State
- 5:10bill that was meant to regulate really large AI models was
- 5:13vetoed by Governor Newsom. But maybe there'll be a national
- 5:16regulation, maybe there won't. Where do we think Trump leans
- 5:18there? Markets and MA is a huge 1.
- 5:22The mergers and acquisitions, the roll ups of small tech
- 5:25companies have been severely depressed under the Biden
- 5:28administration. Lena Khan, his Federal Trade
- 5:31Commission chair is sort of openly anti tech and anti
- 5:33acquisition. Will that change under Trump?
- 5:37And then finally, I think you just can't leave out the old
- 5:39chaos, Black Swan madness of the Trump administration.
- 5:43While shit seemed to happen on a weekly basis, January 6 being
- 5:47the real cherry on top of that Sunday.
- 5:49But what? What kind of crazy stuff could
- 5:52happen? And how does that affect, you
- 5:54know, tech and AI? Obviously, hopefully not crazy
- 5:57enough to lead to, you know, bigger negative consequences,
- 5:59but as. You say nuclear.
- 6:02Well, nuclear or just Cold War, you know, Cold War 2, you know,
- 6:06maybe like. A pandemic or something?
- 6:08Yeah, like a crazy global pandemic.
- 6:10Yeah, I think that's that was a good Black Swan.
- 6:12Not Trump's fault though. I think we can we can agree.
- 6:14Maybe let's talk energy. How do we make enough energy to
- 6:18power GPT 8 or whatever's going to come in the next 4 years?
- 6:22Again, I think to frame this up, we're already seeing a loosening
- 6:24a nuclear policy and regulation. Texas is building more solar
- 6:28panels than the rest of the country because it's easier to
- 6:30build there. How are we going to make enough
- 6:31energy? And is Trump good for energy or
- 6:34not? Well, I will say that the bid to
- 6:37use nuclear to power that Amazon data center was rejected by the
- 6:43Federal Energy Regulatory Commission under Biden.
- 6:46So I think we'll see what happens.
- 6:48But I could definitely see Trump administration being pro nuclear
- 6:52and reversing some of these decisions to reject efforts by
- 6:56big tech to use nuclear power. Restart these nuclear reactors.
- 7:01Hopefully, you know, I, I think more likely that not he's
- 7:03friendlier to nuclear, but I don't think we we know for sure.
- 7:07I mean, I think nuclear is a tough one because everyone's
- 7:10afraid of meltdowns, you know, Chernobyl, yadda yadda, yadda.
- 7:14Even though I think you can objectively say nuclear hasn't
- 7:16killed that many people. Nuclear energy, for the record,
- 7:18not nuclear weapons, right? It's just very scary to people.
- 7:21It's just got this fundamentally scary element to it, I think,
- 7:24like. Populist.
- 7:24That's not a populist, yeah. It's not a populist energy
- 7:27source and Trump is nothing if not very sensitive to populism.
- 7:30I think solar is a much more interesting one.
- 7:31I just keep seeing more and more data about how much cheaper
- 7:35solar's gotten in the last couple decades.
- 7:37It's now basically cheaper than any new source of energy.
- 7:40And Texas is rocking so. Right.
- 7:42I mean, it is like, just deregulate, Don't tell people
- 7:45what to do with their houses, and you'll get a lot of solar.
- 7:48Yeah. And so Texas.
- 7:49Texas solar is really the argument that, you know, a
- 7:52deregulatory environment is perhaps better for the climate
- 7:56than people who vibe more with climate change, climate change
- 8:00being an issue. Like Texas is probably the most
- 8:03impactful state in the union in terms of progressing our clean
- 8:05energy future because you can just throw up solar panels in
- 8:08like a couple months there. And you can't do that anywhere
- 8:11else in America, certainly not California.
- 8:13And so arguably, Texas is having a more positive impact on
- 8:17America's transition and clean energy than any other, you know,
- 8:20major state government or real than the federal government,
- 8:22certainly. I hate that the first one is pro
- 8:25Trump. That's our bias.
- 8:26I showed it right there. I mean, one thing.
- 8:28I mean, you know, the yeah, I do think it's not surprising that
- 8:33on the measurement of unrestrained energy production
- 8:37without much regard for environmental impact, who's
- 8:41going to win any Republican, generic Republican?
- 8:44I think is is sort of a traditional answer there.
- 8:48I mean, I think to your point, we can caveat all this with
- 8:50Trump changes his mind all the time.
- 8:52He did Operation Warp Speed and now he basically wants to ban
- 8:55vaccines. So like, you know, no guarantees
- 8:58that were their stated situation is where we'll end up with solar
- 9:01panels. But I think trying to be a
- 9:03little optimistic in the midst of sort of darkness of this
- 9:06experience, that maybe this will be really good for clean energy
- 9:09in the United States. And obviously there's also just
- 9:11the risk of like favoritism where it's like Amazon maybe
- 9:15never gets their nuclear contract approved, but Elon gets
- 9:19three of them. So that that'll be another
- 9:22issue. Totally in play.
- 9:25Chips. This should be a winning issue
- 9:27for Biden. He passed legislation called the
- 9:31CHIPS Act. Could not be more overtly
- 9:34aggressive on this issue. Trying to make sure that, you
- 9:37know, America is not dependent on China really to produce chips
- 9:41that power, you know, all our devices.
- 9:44Of course, today, you know, Taiwan, which at the end of the
- 9:49Biden administration is not China, is, you know, through
- 9:52TSMC, a major producer of silicon for the United States.
- 9:56I don't know what. Yeah.
- 9:57What do you think? Is Trump a gain or loss on the
- 10:01chips front, James? Trump in the last few months of
- 10:04the campaign was saying on the trail that the CHIPS Act was so
- 10:08bad that the only way to get these companies or the best way
- 10:13to get these companies back to manufacturing in America is to
- 10:17put a bunch of tariffs on their imports so the they're forced
- 10:21back to America. So I'm a little skeptical that
- 10:25he's going to be sticking with the the Biden strategy here,
- 10:30even though it seems to be succeeding in the US.
- 10:33But I could see him changing his mind to once kind of the clarity
- 10:38of the issue is in front of him. Yeah, I think like solar where,
- 10:44I'm sorry, energy in general, where we were sort of like,
- 10:46yeah, based on what they're saying, this policy could be
- 10:49pretty good for energy production.
- 10:51I think we have to go based on what they're saying to James's
- 10:53point, and it's not good. Like, you know, they hate the
- 10:57chips acts. Mike Johnson, the speaker of the
- 10:59House, said they're going to try to repeal it even though his
- 11:01like Co speaker at an event was like, actually that was really
- 11:05good for my district and I'm a Republican.
- 11:07So can we not do that? So bizarre.
- 11:09It's industrial policy. Like I don't understand.
- 11:11Tariffs. Tariffs are a terrible way to
- 11:13deal with this problem. I mean, I fundamentally will get
- 11:17this in the Black Swan section, but I think that Trump makes it
- 11:19much more likely that China will try to take over Taiwan in the
- 11:22next 4 years, which would be the worst possible thing for chips
- 11:26in the global on a global scale. Obviously, the reason we need to
- 11:30bring semiconductor, semiconductor production here is
- 11:33because of the fear that we don't have access to Taiwan.
- 11:37As long as we do, we're in a pretty good spot.
- 11:40It's obviously up for a debate, but I pretty strongly believe
- 11:43that Trump is enabling China and Russia.
- 11:46But let's just focus on China right now and they're sort of
- 11:49leaders to act in a more dictatorial fashion, dictatorial
- 11:51fashion. And clearly the number one thing
- 11:54Xi Jinping wants to do is retake Taiwan.
- 11:59They have the US gate set right? Haven't they said like 2046 or?
- 12:02Something they're doing like practice blockades in the ocean
- 12:06around the island. They're like, we're just going
- 12:07to throw like 9 battleships in a little circle around Taiwan just
- 12:11to just to feel out the waters. So it's not, it's not looking
- 12:16good in my opinion. I don't know.
- 12:17I'm pretty I'm pretty pessimistic about the future of
- 12:20the Taiwan China conflict. All right.
- 12:23We're moving on to regulation. I think it's pretty obvious that
- 12:27the VCs and Elons of the world who helped get Trump elected
- 12:32believe that Trump will just won't regulate.
- 12:35You know that. What is Doge Department of
- 12:39Government? Department of Government
- 12:40efficiency. Exactly.
- 12:41We're going to cut everything. Why would we add any regulation
- 12:44to get in the way of AI? Let it RIP.
- 12:46Baby, we should talk about what is Elon trying to achieve from a
- 12:50regulatory perspective because he was supportive of SB1047, the
- 12:57California bill that government Governor Newsom vetoed this
- 13:02year. So I am a little confused.
- 13:04Maybe you guys have a better sense of what Elon prefers from
- 13:08a regulatory. Perspective he prefers he be the
- 13:11one to do it you know, and everyone else be regulated is
- 13:14my. Checks out.
- 13:16Didn't he? Didn't he suggest that SpaceX
- 13:18executives take be part of the Pentagon?
- 13:20Yeah, I think that was a story that came out today.
- 13:24It's seems like it's what's good for Elon is apparently good for
- 13:27America. But yeah, I mean, Elon is
- 13:29extremely worried about sort of, you know, the big existential
- 13:33risks of AI. That's why he helped create open
- 13:37AI in the 1st place. Now open AI is another source of
- 13:40worry for him. So he's been supportive of rules
- 13:44to try and make sure we don't create some super intelligence
- 13:49while at the same time now building his own foundation
- 13:53model company with XAI. Yeah, The interesting thing
- 13:56does, does Elon look for regulation when he's behind in
- 13:59an industry and then, you know, where he's ahead?
- 14:02He looks for deregulation. I think if you look at even even
- 14:06things like SpaceX or Tesla, you know, he's advocating for
- 14:10deregulation because and potentially he would benefit
- 14:14from, you know, electric car subsidies being revoked because
- 14:19he's so far ahead that other companies are more reliant on
- 14:22those subsidies. Similar for SpaceX, you know,
- 14:24he's so far ahead that anything that allows him to launch more
- 14:27rockets faster is good for Elon. AI not not as clear because if
- 14:32he, you know, if, if open AI and anthropic and others just kind
- 14:35of go at the pace they're going, you know, he's he's a little bit
- 14:38behind other AI related regulation, though, I guess
- 14:42would be something maybe we should discuss around wokeness
- 14:45of the AI or free speech. Clearly, that's sort of racial
- 14:51bias and stuff is not going to get regulated.
- 14:53I also think, you know, Hollywood has a potential, you
- 14:57know, strong copyright interests in AI, and they seem sort of out
- 15:01of the fold in a Trump administration.
- 15:03So you could have imagined a Paris administration sort of
- 15:06trying to do something to help protect copyright holders, and
- 15:09maybe that doesn't happen. I also think it's interesting to
- 15:12your point on the the wokeness of AI issue, like, you know, JD
- 15:17Vance kept bringing up this like censorship of social media as
- 15:21like a terrible crime of a Democratic administration,
- 15:24right? And obviously, I think we all
- 15:26know that the current top of the line AI models, GPT and, and
- 15:30Claude from from Anthropic are are, you know, there's a lot of
- 15:35quote UN quote guard rails around what you can ask it,
- 15:38right? You can ask it if you asked to
- 15:40build like a pipe bomb or whatever, it's not going to tell
- 15:42you, right? And then there's a lot of other
- 15:44sort of I don't, I don't know if you would categorize them as
- 15:47woke, but you'd say safety related sort of actions taken by
- 15:53those companies. And my question is like, is
- 15:56someone like JD Vance in the VP slot going to go crazy on these
- 16:00companies that are, you know, regulating the speech of AI
- 16:03quote, UN quote and try to open up these models and make them a
- 16:07little wackier and maybe a little less, less safe when you
- 16:10talk to them? Yeah, the weird part about all
- 16:13this is there is no regulation on this right now.
- 16:15It's all sort of internally regulated by the companies.
- 16:19Yeah, an AI has a set of things they don't want ChatGPT saying
- 16:23to you. And it's true of of X AI of
- 16:27Grok, I believe when they launched their image model, you
- 16:31could ask for literally any copyrighted characters.
- 16:34You could ask for Mickey Mouse holding assault rifles or yeah
- 16:39it. Was a fun moment.
- 16:41Celebrities, yeah. And I think Elon himself also
- 16:45rolled that back, whether that's, you know, due to
- 16:47existing copyright law or, you know, their own internal desire
- 16:51to not have that be what users can make in their in their tool.
- 16:55It wasn't really some unique new AI regulation that that had him
- 16:59do that. Importantly, the Biden
- 17:01administration introduced an executive order that, in the
- 17:06words of the Andreessen Horowitz crew, regulated math.
- 17:10Or made me, you know, try to make math illegal.
- 17:12It's like if you're doing a lot of math, if it's big enough, we
- 17:15need to register with the government and we're going to
- 17:17keep track of it. And so that that was mostly
- 17:20monitoring, but it certainly was a sign that, OK, government is
- 17:24going to be much faster this time around than when with
- 17:27social media to sort of police the development of AI.
- 17:31So there there's a big question of does the Trump administration
- 17:35pick up the ball and say, oh, that was a good idea, We should
- 17:38keep keep doing stuff there? Or do they say, oh, no, like our
- 17:42tech people, they don't really like that.
- 17:45And and I think part of what we're sort of dancing around
- 17:47with Elon is like some of the tech people, you know, Marc
- 17:50Andreessen, Ben Horowitz, they endorse Trump.
- 17:52They don't want AI regulation. But then you've got Elon who
- 17:56seems to be very worried about it.
- 17:58So it's not clear that there's actually a Silicon Valley Trump
- 18:01supporter consensus. And so therefore, we don't know
- 18:04exactly what is going to what Trump is going to believe in
- 18:08terms of how much the either executive orders or pushing
- 18:13legislation needs to come along in AI.
- 18:16I just really, I think this issue is gonna come up around
- 18:21job losses and job replacement at some point, Like all like
- 18:25half of American politics is all about jobs in the end.
- 18:27And I think that if and when we get to the point where there's
- 18:30meaningful job losses among a certain sector of workers, some
- 18:33party is going to step in and be the anti AI party for that
- 18:38aggrieved group of workers. If we get self driving trucks
- 18:42pretty soon and the 1.2 million truck drivers are losing their
- 18:46jobs, like that's probably a Republican anti AI group, right?
- 18:50Like you know, F the AI that are driving our cars.
- 18:53Like we need a regulation that mandates like there's always a
- 18:55human behind the wheel or something.
- 18:57And then on the other side, like for, you know, Democrats, if it
- 19:01turns out it's a bunch of like college educated, like middle
- 19:05management, white collar workers, you know who, you know,
- 19:08legal associates or something, we're all getting replaced.
- 19:10And there's like massive job losses and those markets, I feel
- 19:13like the Democrats might step up and start being the anti AI
- 19:16party or the, you know, the, the white collar job party to
- 19:20protect against AI. So I just, my prediction is that
- 19:23at some point in the next 4 years, there's going to be a
- 19:25meaningful anti AI backlash builds around some sort of job
- 19:29losses. And I think it's really about
- 19:31what job losses come quicker or what, what group of immuno
- 19:35employees starts being more and more nervous about getting
- 19:38replaced by AI. Because if you look at the
- 19:40public polling, people are pretty nervous about AI already.
- 19:43And I think that's just going to keep escalating on a, you know,
- 19:46monthly, quarterly, yearly basis.
- 19:48Couldn't be more opposed to that.
- 19:51But yeah, I agree with you that both parties seem likely to take
- 19:56up the mantle. I mean, I guess it's going.
- 19:58To be a popular viewpoint like. And sympathetic to the worker,
- 20:02you know. Like.
- 20:03America's version of like a safety net is like you should be
- 20:06able to get a job. And so if we're not going to
- 20:09give, you know, much stronger social benefits to people, then
- 20:14we're going to have to sort of take a heavy-handed approach to
- 20:17making sure that they can get jobs.
- 20:20I would much prefer to say, as you know, we attack jobs and
- 20:24find efficiencies, give people more healthcare coverage and
- 20:27safety net benefits than these sort of backwards looking job
- 20:31protections. I guess maybe a counter take to
- 20:35that would be that one of the main lessons of the last
- 20:39election was that people seem to hate inflation a lot more than
- 20:43they hate unemployment. So we'll maybe if AI is somewhat
- 20:47deflationary, you know, in. Terms of eggs, can your eggs can
- 20:51be cheaper all right, like. Productivity going up.
- 20:54There's robots, robots, farmers now picking those eggs.
- 20:58Yeah, I don't know, Maybe then it'll be hard to justify some of
- 21:03those regulations if you know you're seeing the the benefits
- 21:07of of a bit of the cooling of inflation.
- 21:10And I just don't, I really disagree with the idea that
- 21:12people hate inflation when they hate unemployment.
- 21:14I think it's all about the order of magnitude of both of these
- 21:17things, right? Again, Like, yeah, people
- 21:18dislike 9% on, you know, inflation much more than they
- 21:22dislike like 4 to 5% in unemployment.
- 21:24But if you go back to the great financial crisis in OA, like
- 21:27unemployment is the reason George Bush and John McCain got
- 21:30completely incinerated in that election, right?
- 21:32It wasn't inflation. It was like job losses and the
- 21:34economy crashing everything. So I just don't, I think that if
- 21:385% of American workers lose their jobs due to AI in the next
- 21:41four years or something or 10, like it's people are going to
- 21:44hate that. That would be my prediction.
- 21:48If there were a prediction market for for that I would I
- 21:50would throw some throw some Dogecoin on it.
- 21:52I think one of the key reasons we all oppose Trump among many
- 21:56is just like he's a chaos agent. Who knows what's going to
- 21:59happen. Like America is actually great
- 22:01right now. Like why blow up the status quo
- 22:04in some ways? Obviously we could make
- 22:06continued incremental improvements, but just like,
- 22:08yeah, chaos is not necessarily a good thing.
- 22:11But but yeah, so I I think predicting what's going to
- 22:14happen in Trump administration without considering sort of
- 22:17insane things going to going to happen makes him seem sort of
- 22:20more normal than he is. We've already touched on Taiwan.
- 22:23And of course, if China invades Taiwan, that hurts
- 22:26semiconductors. Are there other scenarios you
- 22:29can imagine where having Trump as president and this continued
- 22:33improvement of AI causes causes some weird problem or benefit?
- 22:40Yeah. I mean, I think we should talk
- 22:41about what the world would look like under Trump if we achieve
- 22:45AGI in the next couple years. Maybe that's an early timeframe,
- 22:49but that's seems like what Sam Altman is offering.
- 22:52Well, you know, some of these people think AGI will solve
- 22:56democracy, so maybe we can get rid of Trump early.
- 22:59We plug it in the computer. It's like.
- 23:01Why do we need democracy? Or just, you know, it's like
- 23:04solves all these optimization. I'm joking.
- 23:07I honestly don't believe that. And Dario at Anthropic wrote
- 23:10like a review of how he thinks like AGI could move us forward.
- 23:14And I think he was also skeptical that democracy would
- 23:17be 1. But yeah, and recently said on
- 23:21an interview that he expects us to kind of just blow by the AGI
- 23:28sort of goal post and then society not to change that much,
- 23:33similar to maybe how we don't feel like society changed that
- 23:37much by passing the Turing Test. So maybe it just is irrelevant.
- 23:40But it could be pretty interesting if that's not the
- 23:44case and AGI totally upends society in some ways.
- 23:49Max already mentioned the kind of job loss scenarios.
- 23:52Is there other anywhere else that we might see AGI like
- 23:55completely change American Society?
- 23:58What about AGI, you know, what about AI mayors and AI governors
- 24:02and sort of AI actually taking positions of power in our
- 24:05government, right? I mean, like, I think that
- 24:07there's some candidate in in, I forget, in Texas or somewhere
- 24:11else who said he's going to like put every decision that he has
- 24:13to do as mayor through ChatGPT. Like, are we going to see that
- 24:16become a more meaningful part of the actual functioning of
- 24:20American government that like, decisions are being filtered
- 24:23through AGI or AI agents are carrying out administrative
- 24:25roles? Like I think it's possible.
- 24:28Possible one thing I think I would flag though is if Trump,
- 24:32whoever happens to be president, if there is this big inflection
- 24:36point in AI, while we, you know, fantasize about downsides, could
- 24:40actually be beloved. You know, it's like, oh, all of
- 24:43a sudden we're making way more progress on healthcare without
- 24:47spending government money. All of a sudden we have, you
- 24:50know, these improvements, things are going great.
- 24:53There's a world which like, especially if his message is
- 24:56like, I'm going to let capitalism run wild and
- 24:58capitalism delivers AGI and AGI actually is good for people.
- 25:03You know, the president, you know, sort of gets to benefit
- 25:06from whatever, whatever is happening.
- 25:08And, you know, if there are huge improvements in the next couple
- 25:12of years, it could just sort of boost people's impression of a
- 25:16Trump presidency. I do think that's possible.
- 25:20I could see maybe Max what you're saying.
- 25:24There have been people who have tried to, I think, use ChatGPT
- 25:29as their defense attorneys. At some point, like does ChatGPT
- 25:32just become better than your average government provided
- 25:36attorney? Maybe then it starts to be an
- 25:39interesting legal question of whether you're allowed to do
- 25:41that or not. It's hard to say that whether
- 25:43Trump would be supportive of any of this.
- 25:47All right. So a quick look back, I think
- 25:49energy, we think, you know, Trump could be good for creating
- 25:53the energy production we need for AI, though of course could
- 25:56have any sort of terrible impact on climate change chip
- 26:00production. I think the risk that he
- 26:02escalates tension with China and therefore hurts our ability to
- 26:06get chips from TSMC and Taiwan is probably a negative.
- 26:09He's also somewhat a tariffs, he's somewhat opposed to the
- 26:14Chips Act, which is actually going well.
- 26:15And it seems like some of these plants are actually working.
- 26:20Regulation seems like he has competing points of view from
- 26:25his supporters. So it's unclear who he's going
- 26:27to listen to or how that plays out and or whether they're
- 26:29sophisticated enough to do it. We didn't really dig into
- 26:33markets, but I think it's clear that right now the stock market
- 26:37is bullish about Trump. If there's not a lot of MNA
- 26:40review that there could be, you know, if antitrust is loosen up,
- 26:43there could be a lot of acquisitions and then it's very.
- 26:46Good for MNA, yeah. Yeah, so good.
- 26:48Good, good for M&A, which I mean it could be bad for development
- 26:53if you think a lot of startups is, is how we make improvements.
- 26:56But yeah, so, so I I think it's. I would add one final footnote,
- 27:03which would be the final Trump Black Swan is he refuses to
- 27:07leave office and we have a constitutional crisis slash
- 27:11downfall of Americans democracy scenario in four years, Which I
- 27:14don't think there's a lot to say about that other than it would
- 27:16probably be quite bad for everything across the world, not
- 27:19just AI. But I think that that's always
- 27:21the Black Swan you have to throw in there with Trump because he's
- 27:23clearly shown an intolerance for being removed from the White
- 27:26House. Great.
- 27:31Well, I mean, I, I agree with you 100%.
- 27:34Like to be clear. And I think generally the point
- 27:37is that you have to be alive and a functioning society to benefit
- 27:41from AI and that if you are not, you won't.
- 27:44And so that is why I think many of us thought that that was the
- 27:47number one issue of the the markets and whether anybody can
- 27:52acquire anybody or not. Yeah.
- 27:54All right. So now we're going to move on to
- 27:56our next segment here, a little game we'd like to call
- 28:00overrated, underrated or properly rated.
- 28:04The goal of this podcast is to sort of assess where we are
- 28:07today in AI, how various companies are doing.
- 28:10And I think what's interesting, what James and I have discovered
- 28:14as startup founders, is there's often quite a lot of data
- 28:16available to anyone who signs up for the right websites about how
- 28:22companies are actually doing. And for these companies, a lot
- 28:26of which we've talked about in the AI startup space, the key
- 28:30indicators of success tend to be either revenue or users and
- 28:35usage. And you can find this kind of
- 28:39data for both categories. You can find it for companies
- 28:43that sell to other businesses, which are more revenue focused
- 28:46and consumer companies, which are more about usage and uptake.
- 28:50And so I thought it would be fun to kind of walk through about 10
- 28:55top AI start-ups, talk through their actual data that we were
- 28:58able to collect from various sources and then walk through if
- 29:02we think, given this data, if the company is overrated,
- 29:05underrated or properly rated based on fundraising and hype
- 29:09and so on. And just to lay this out up
- 29:12front, we pulled the revenue data from the information which
- 29:15we believed to be fairly reputable source with respect to
- 29:18company revenue, very reputable. But it's, it's clearly a chart,
- 29:21right? So it's yeah, OK, Some of it's
- 29:23directional, very reputable. Yeah.
- 29:26And then the usage data is pulled from two sources.
- 29:29One is called similar web, which I would consider to be the best
- 29:33in class tracker of website usage for more website driven AI
- 29:38companies. And then data dot AI is probably
- 29:43the best for app data. So iOS and Android app usage.
- 29:49Again, none of this stuff is like perfect within, you know, 5
- 29:52or 10%, but it tends to be very good about order of magnitude.
- 29:55Is this a company with millions of users or 10s of millions of
- 29:59users or thousands of users? And it also tends to be good
- 30:01directionally. Is this company growing flat or
- 30:05shrinking, which I think is very valuable information.
- 30:08Especially in AI world where companies will have raised that
- 30:11billion dollar valuations a year ago and now people just sort of
- 30:15move on from them. And the only reason you might
- 30:18know why is if you look at their growth charts and they're
- 30:21they're reversing. Yeah, absolutely.
- 30:23So I'm going to pick 10. We're going to go through
- 30:27debate. Overrated, underrated, properly
- 30:29rated. I will frame it up first with
- 30:32the big kahuna, Open AIIE, ChatGPT.
- 30:36According to the information, Open AI is currently making over
- 30:40$4 billion with AB in annualized revenue and I was able to pull
- 30:45the data both on app usage and website usage for chat, TPT and
- 30:49Open AI. Open AI has over 25,000,000
- 30:53monthly users on it's app in the United States and has over
- 30:57284,000,000 monthly users worldwide on it's website.
- 31:02I think you could pretty safely say it probably has near 1/2
- 31:07billion monthly active users if you were to add in the global
- 31:09app data. USIOS 25,000,000 Yeah, and I
- 31:15think most interesting is the growth curve.
- 31:17According to basically all third party sources, Open Eye has
- 31:20grown almost 50% just in the last three months, meaning it's
- 31:23like 1.5 X bigger than it was three months ago.
- 31:26That could be the back to school boom as we've discussed in
- 31:29previous episodes, but it has also grown over 2X year on year
- 31:33in terms of usage, as in it has twice as many people using it as
- 31:36it did a year ago. So.
- 31:38Huge growth. Yeah, so overrated, underrated,
- 31:41properly rated. This is obviously the highest
- 31:43rated startup in the AI space, but I'm interested in your guys.
- 31:46Take James. Interested in your take.
- 31:49Overrated, underrated, Properly rated.
- 31:52OK. So $157 billion valuation,
- 31:56there's maybe some question marks around that valuation in
- 32:01terms of how company, how investors get paid back and, and
- 32:05things like that. I would say probably properly
- 32:08rated. It does seem to be the technical
- 32:11leader in the, in the foundation model space.
- 32:16The one thing that I found really interesting is their CFO
- 32:19said recently, 75% of revenue comes from consumers, so mostly
- 32:23ChatGPT subscriptions. The reason I would say properly
- 32:26rated is it's a huge amount of revenue from consumer
- 32:29subscriptions and very few companies have that level of
- 32:33consumer subscription model business model.
- 32:36But I do, I think that what's interesting is that the core of
- 32:40the company doesn't seem to be much of A consumer company.
- 32:43They they did hire Kevin while. So maybe there's some efforts
- 32:48internally to bring in kind of consumer entrepreneurs and
- 32:51operators. But I think, yeah, I'm a little
- 32:53nervous about whether they can actually capture the value of
- 32:56having this massively successful consumer product sort of
- 33:00accidentally inserted into a research lab.
- 33:04I'm rare enough and ready to go. Go baby.
- 33:07I I think it's overrated. I mean, I've written a whole
- 33:10column about how it is overvalued.
- 33:13You know, I, I think certainly the revenue growth, which is
- 33:16what Max you're focusing on, is astounding.
- 33:19And if you look at that growth and compare it to Google and
- 33:22Facebook, I guess valuation makes sense of sense.
- 33:24Yeah. But you know, it's burning like
- 33:27$5 billion this year. It's going to burn probably more
- 33:30next year. And all that money is being
- 33:33spent creating models that are getting commoditized.
- 33:37So I think that's a big challenge.
- 33:38But the thing I'll add besides the piece you can go and read is
- 33:42that I think it's also overrated in that people in Silicon Valley
- 33:46are excited about Perplexity and Anthropic and everything else.
- 33:49And there are lots of other sort of competitors that people don't
- 33:52know. And so the public massively
- 33:55rates open AI, Chachi PT. That's what they know, that's
- 33:58what they're using. And maybe that means that open
- 34:00AI runs away with it. But if you were just talking to
- 34:03the consumer, they absolutely overrate the quality of open AI
- 34:07relative to what's available on the market.
- 34:10Yeah, 100%. I'm not going to weigh in too
- 34:13long, but the growth is freaking nuts.
- 34:16I think given that the models have not improved that much in
- 34:18the last year, we can all say 2X growth year on year and 50%
- 34:23quarter on quarter growth is absolutely monstrous.
- 34:25So to me this kind of re rated me to be in the more properly
- 34:28rated zone. If it's multiple AT35X revenue,
- 34:31like that's just crazy growth given the models haven't even
- 34:34gotten that much better. If we actually get a better
- 34:36model and you continue to see even faster growth in this, the
- 34:41valuation is completely justified.
- 34:42So I'm, I'm sort of fascinated to to see if this growth group
- 34:45continues. All right, moving on to our next
- 34:47one. We'll go with the classic
- 34:49pairing with Open AI Anthropic, makers of Claude dot AI.
- 34:55Anthropic is making about 100 million in revenue, according to
- 34:58the information. It has almost 10 million monthly
- 35:01users on its website, not super meaningful app usage, you know,
- 35:08maybe a half million to 1,000,000.
- 35:10And it grew 20% last quarter in contrast to Open A is 50% growth
- 35:16in the last quarter. So less than half as fast growth
- 35:19of Open AI. So this time I'll kick it off
- 35:21with you, Eric. Overrated, underrated, Properly
- 35:24rated. I'm going to say underrated.
- 35:31I think interesting one. One thing that I think is
- 35:34interesting is just, if Trump really loosens the M&A
- 35:39environment and Amazon is in a sort of dire position in the AI
- 35:45race, does Amazon acquire Anthropic?
- 35:48I think it becomes much more likely all of a sudden.
- 35:50And, and so I, I think it's the, the value of Anthropic is, you
- 35:56know, the amazing technical staff, they keep recruiting
- 35:59defectors from open AI. Obviously I believe in the power
- 36:04of revenue and money coming in and open AI is dominating it.
- 36:07But in many ways, Anthropic has sort of the same thing.
- 36:11And if this is a great category, there's going to be more than
- 36:14one player and they'll figure out how to how to get customers.
- 36:19James, I'll agree with Eric, underrated.
- 36:23I think they have a diversified strategy, most of their revenues
- 36:27coming from, not from consumers, right.
- 36:30So they have like every reason to go after enterprise adoption
- 36:36and starting with what everyone basically agrees upon in terms
- 36:40of the comparison to open AI that their coding models are,
- 36:45are better. Maybe that will change.
- 36:47And you know, who knows what happened, what open AI has
- 36:50internally, but they seem to be outperforming open AI in that
- 36:56very key market of, of, of developer adoption and encoding
- 37:01capabilities. So I think underrated for that
- 37:05reason. Yeah, I will quickly say, based
- 37:09on this data, I think super overrated.
- 37:12Yeah, based on the data, yeah, we're just ignoring the data.
- 37:15I mean, you guys can just say whatever you want.
- 37:16The vibes. For the record, before I pulled
- 37:19this data, I would have been where you are, which is like
- 37:21anthropic. The model's dope.
- 37:23Everybody loves it. Developers.
- 37:24It's like Silicon Valley, darling.
- 37:26It's the cool, cool, artsy version of open AI.
- 37:29OK guys, Opening Eye is making 40 times as much money and has
- 37:3325 times as many users as Anthropic and is growing more
- 37:38than twice as fast right now and they're currently valued at like
- 37:411/3 as much. It's absolutely cuckoo for Cocoa
- 37:45puffs. It makes no fucking sense other
- 37:47than as an M and a play. So I think Anthropic's great.
- 37:50I love the model, but based on this data, Anthropic's valuation
- 37:53is absolutely ludicrous in comparison to Opening Eyes.
- 37:56So I'll say super overrated. Let's keep going Mid Journey,
- 38:00the darling of the image generation universe.
- 38:04Now I will caveat this with Mid Journey is the one consumer
- 38:07company where data is kind of hard to come by because they
- 38:09kicked off on a discord, meaning a huge quantity of their usage
- 38:12is on a Discord. Server as far as we know they
- 38:14haven't taken outside investment right as.
- 38:16Far as we know they haven't taken outside investment.
- 38:19They I will give three pieces of data here. 1 is the annual
- 38:24revenue of mid journey according to the information is $200
- 38:27million or more and growing. I was able to get data on their
- 38:33web usage of midjourney.com which is 4 million monthly
- 38:38active users and growth looks pretty flat on the website as in
- 38:42not growing, not shrinking, not much.
- 38:45Either way flat. So James Mid Journey.
- 38:50Overrated, underrated, properly rated.
- 38:53I think overrated at this point. I think they were clearly
- 38:58pioneers in this image generation space and they had a
- 39:03nice advantage. You know, people probably a lot
- 39:06of this revenue is, is recurring revenue from subscriptions that
- 39:10they probably acquired a while back.
- 39:12Like you said, maybe not so much growth recently.
- 39:16I think that the image generation model space has
- 39:18gotten more commoditized. There's really good other models
- 39:21out there, flux image generation being one of them that gets
- 39:25high, high, high rate in terms of realism and and accuracy.
- 39:31So just a commoditized space and I'm not sure that the the
- 39:35consumer revenue side is going to keep growing from from mid
- 39:37June. Eric, I, I agree with everything
- 39:40James said, but I'm going to say underrated for the very reason
- 39:44that they just didn't take venture capital.
- 39:45It's all like house money. They're making, they're making
- 39:48money and they're pocketing it. And then so like, even if they,
- 39:51like, aren't going to be a sort of grow, grow, grow story where
- 39:54they're going to be the next, you know, public $100 billion
- 39:58company, and still they made a killing along the way.
- 40:02So I say kudos to them. Yeah.
- 40:05Underrated. Yeah, I think I see both sides
- 40:07of the debate. So I'm in I I support both of
- 40:11your takes. Well, then you're properly
- 40:12rated. You're.
- 40:13Properly rated, yeah, I think properly rated.
- 40:15I think the lack of growth is dispiriting.
- 40:17But to Eric's point, it's an absolute ATM machine for the
- 40:20founders, so who cares, I guess. All right, moving on, Jasper
- 40:25frequently ridden a hyper roller coaster, been high, been low.
- 40:29You guys can weigh in on where it is in the middle right now,
- 40:31according to the information making around 60 million in
- 40:34annualized revenue. It has 600,000 World Wide Web
- 40:39views in the last month according to Similar Web and
- 40:42it's basically flat on growth 600,000, not 600 million.
- 40:47Yeah, yeah. Overrated.
- 40:49Overrated over, over, like overrated.
- 40:53All right, James, just no, no, I'm saying, like, does the
- 40:55company even matter anymore? It was like, you know, we were
- 40:58fighting to get them at our March Shrubal Valley last year,
- 41:02and now I feel like I didn't hear anything about them.
- 41:04OK, James. I guess I maybe I'm calibrating
- 41:09to what I thought was a company that was already to Eric's
- 41:14point, sort of like sort of it's, it's time had passed.
- 41:20But I guess based on where I was coming into this podcast and and
- 41:24the, the 60 million revenue figure, like I think that's
- 41:27maybe they're a little underrated.
- 41:29You know, given nobody, nobody's really talking about them, but
- 41:32they, they have a pretty substantial revenue base.
- 41:35If it's almost like more than 1/2 of anthropic, I don't know,
- 41:38that's pretty interesting. Maybe that's outdated data or
- 41:42something from the information. I'm not sure how recently that
- 41:4560 million it was reported, but if it's if it's kind of in that
- 41:4960% of anthropic, I would I would be a little bit longer
- 41:53Jasper than I was coming into this.
- 41:56All right, yeah, I think Flack grows the step.
- 41:58I, I, I would only say properly rated because I think folks like
- 42:02Eric have already danced on their graves a few times.
- 42:04So properly rated is fairly negatively.
- 42:08OK, let's do SUNO, the music generation darling that lets you
- 42:14create cool online songs using generative AI.
- 42:18They are at 40 million in ARR. They are at around 12 million in
- 42:23World Wide Web views or web users.
- 42:25I'm sorry, uniques over the last month.
- 42:27And I think most interestingly, according to similar web, they
- 42:32appear to be shrinking a little bit over the last three months.
- 42:36Again, caveat the status from a third party source, but I think
- 42:39we can safely say they're not growing and they might be
- 42:41shrinking in the last three months.
- 42:44So, yeah, Eric, or sorry, James, yeah.
- 42:46I had some questions on this. So did you, you mentioned their
- 42:49mobile app adoption that that they launched that like I'm
- 42:52wondering if some of the web traffic is kind of?
- 42:55Cannibalized by the mobile app, yeah, the mobile app's quite
- 42:57small in comparison the website right now.
- 43:00So I, I don't think it's enough to offset the shrinkage which
- 43:04looked like 15% quarter on quarter order on similar web.
- 43:08But yeah, there's probably some cannibalization of the web app
- 43:12by the mobile app. But again, I just don't think
- 43:13it's enough to offset. The mobile app isn't growing
- 43:16super fast. Or something.
- 43:17It's growing fast, but it launched like two months ago.
- 43:19I mean, so basically doubled month on month from like 100K to
- 43:22200K, you know, So it's not like life changing numbers here,
- 43:26yeah. What is it?
- 43:29James overrated, underrated, Properly rated Suno.
- 43:33I know you love Suno. Yeah, I love Suno.
- 43:35I love the product. They use it a lot.
- 43:37I think they have definitely the best music generation model that
- 43:41I've seen. And they're also like really
- 43:44shipping pretty fast. You know, they ship this pretty
- 43:46cool feature that allows you to take pictures and generate a
- 43:50soundtrack based on the picture. They have these cover features
- 43:55that are allowing you to kind of, you know, sing into you,
- 43:59however you sing, sing version of the lyrics that you want, and
- 44:03it'll kind of put a style onto that music and you can actually
- 44:07save styles now that you're generating from music.
- 44:09So I just think they're very good product company and I would
- 44:15say the 40 million is very impressive.
- 44:18ARR So I'm going to say underrated.
- 44:19I think, I mean, I think everyone knows there's going to
- 44:22be a bear case here of like the legal status of the company, but
- 44:27I could also see that being, you know, taking 10 years to sort
- 44:30out or something or leading to some sort of like Spotify
- 44:34acquisition. So I don't know.
- 44:36I'm I'm pretty bullish on them and I would say underrated.
- 44:41Eric. I'd say properly rated.
- 44:44I mean they were extremely buzzy at Cerebral Valley New York.
- 44:49Their CEO spoke so definitely I am saturated with the hype and I
- 44:58have churned. You know, I didn't I didn't find
- 45:02sort of continue to use to sort of a fun.
- 45:04It's fun to like whip it up and like show somebody, Hey, I made
- 45:07a song about this thing. But I am impressed.
- 45:11I mean they're they're like second on our list in terms of
- 45:14engagement, right? I mean, they're they're ahead of
- 45:17everything. But over AI, yeah, I mean, they
- 45:19are. People are using it and people
- 45:21are using it. Yeah.
- 45:22It's like a clear new use case. Like so much of, you know, AI is
- 45:27like dreaming up use cases or over extrapolating from ChatGPT.
- 45:31And here is an actual generative product that people are using.
- 45:35So properly rated. Yeah.
- 45:37And I think we've talked a lot about like, you know, how far,
- 45:40how long it's gonna take to make Hollywood level movies with AI
- 45:44or even like novels, right? None of that's happened.
- 45:48But like music, we're basically there.
- 45:50Like the quality of Suno is, you know, I think 90% or 95% of, you
- 45:55know, the average pop song or something.
- 45:58So I mean, it's pretty high. Yeah, maybe that's a little
- 46:01high, but at least like it'd be, I definitely think you could,
- 46:03you could just like you'd have to work at it and make some
- 46:08songs that. Hollywood is gonna come for
- 46:11them. I don't know these things are
- 46:12all plagiarism machines, but I. Think if you played songs from
- 46:17someone who just woke up from a coma of the last pop songs of
- 46:20the last year and SUNO songs, you might see some interesting
- 46:24data now. Things like self driving cars,
- 46:26it's like you need to get to 100%.
- 46:28Like 90% is like an uncanny valley where like it sounds like
- 46:33a good song. Like.
- 46:37I'm, I'm gonna go like overrated as a consumer product and maybe
- 46:41underrated as like AB2B product or you know, some sort of like
- 46:44API. Because I think that shrinking
- 46:47at this stage as a consumer company is really, really bad.
- 46:50I think that just whatever whatever spin you wanna put on
- 46:53it of how cool The thing is, I just think shrinking is quite
- 46:56bad or being flat at 11 million even given the level of
- 46:59engagement error. But they are a subscription
- 47:02business, right? So I don't think their revenue
- 47:04is shrinking. Yeah, no, I don't think it is.
- 47:06But I just mean the usage, like if we think it's a consumer
- 47:10product, like tons and tons of people are gonna create songs
- 47:12like for Fun. And that's gonna be I think.
- 47:13That's how they see themselves. I don't think they see
- 47:16themselves as an API company. Yeah, But I agree, I don't think
- 47:19that's how they see themselves, but I think the idea of creating
- 47:22music is very valuable to a million different creative
- 47:25professions. And I think that like, you know,
- 47:28gaming that we work in, I think, you know, having generated music
- 47:31is incredibly valuable and maybe it's an input to actual pop
- 47:34songs and so on and so forth. So I just think that the
- 47:37product, the model is insane. The the I think the revenue is
- 47:40good. To your point, I'm I'm pretty
- 47:42worried about the shrinking consumer.
- 47:43Size I love it is an artist tool where it's like pay for
- 47:46performance and you know, gets all it's money for like we're
- 47:49one of three authors on this song and we get the royalties as
- 47:53a publisher royalties. Yeah, OK, let's let's move on to
- 47:56one of my favorites, Perplexity the the search engine that's
- 48:01going to destroy Google according to Eric Perplexity 20
- 48:06million ARR according to the information.
- 48:08I thought there was recent reporting they put it more in
- 48:10the 30s, so it might be more in the 30s now.
- 48:14Mau of their iOS app in the USI can pull the global Mau, but I I
- 48:20find that I OSUS users to be the most valuable for this kind of
- 48:23product is 900,000 app users under a million and growth quite
- 48:31good 2X in the last eight months.
- 48:34I am also able to pull downloads data from data dot AI and
- 48:39downloads are roughly flat over the last eight to nine months.
- 48:43So kind of an interesting mix of growth narratives there, just a.
- 48:46Clarification there. The Wall Street Journal reported
- 48:49recently that their revenue went to 50 million so.
- 48:52OK. So that's much, much higher than
- 48:54the information reporting. So yeah, up to 50 million in
- 48:56revenue. So if that changes your
- 48:58perception, that's good, good upfront information.
- 49:00So James Perplexity overrated. Underrated.
- 49:03Properly rated. I'm gonna go even despite the
- 49:08fast growth in revenue here, I'm gonna go overrated, given that
- 49:14they're reportedly raising, trying to raise that, I think,
- 49:17an $8 billion valuation. I think it's like one of the
- 49:19happiest companies in Silicon Valley right now.
- 49:24And what's interesting is that they have a ton of competition,
- 49:30including Open AI, which just released this search GPT
- 49:34product. I mean, who knows if any of
- 49:36these can compete with Google on search, But yeah, it seems like
- 49:44it seems like a Google is going to compete and B, like open AI
- 49:48is going to compete aggressively in this space.
- 49:51And, you know, perplexity doesn't seem to have like a ton
- 49:55of sustainable technical advantages or like a super large
- 49:59subscription base currently of consumers or, you know, and also
- 50:03has a lot of potential risks on the copyright and legal side as
- 50:07well. So I'm I'm I'm staying
- 50:09overrated. Eric.
- 50:12It's overrated. I mean, it's just so hyped.
- 50:15It's so hyped. And like, I do think, you know,
- 50:18Google pays a bunch of money to be in search bars.
- 50:20And like, I do think a lot of the search competition is just
- 50:23like being where people go to search and there's going to be a
- 50:27heavy lift winning on quality, but maybe they strike good
- 50:32partnerships or something. Yeah, I think overrated as well.
- 50:36I just think that the $8 billion valuation that they're
- 50:38supposedly raising at, I mean, like for context, like I think
- 50:42their total Mau globally across iOS and Android is maybe like 88
- 50:46million or something like that. And they're worth 8 billion,
- 50:48right? Reddit is currently worth about
- 50:51$20 billion. So we'll say like 2 1/2 X as
- 50:53much right now on the open public markets.
- 50:55And Reddit has about 1.2 billion of the active users.
- 50:59So they're currently being valued at about.
- 51:01And I'm sure they're much more engaged.
- 51:03Too. Yeah, about 40% as much as
- 51:05Reddit with like 11 thousandth the number of or sorry, one
- 51:11100th the number of users like 100, one 100th or 150th of the
- 51:15users. I think the only justification
- 51:19is the 10 to $50 million ARR growth between March and today
- 51:23basically. Yeah, I think the ARR growth is
- 51:25killer, but I just don't think the usage is that killer.
- 51:28OK, I think that's enough companies, a lot of good takes
- 51:30in there. We'll save some more for the
- 51:32future. I think it would be great to get
- 51:33a wrap up and given your guys perception of these companies
- 51:37and the data we've seen so far, what do we think this tells us
- 51:40about the AI industry? What's valuable?
- 51:42Is it revenue? Is it usage, is it growth?
- 51:44Is it a certain type of company? Do you have a more nuanced
- 51:47viewpoint? I'm, I'm fascinated to hear.
- 51:48So, Eric, you want to kick it off?
- 51:51I mean, I think it's notable that and not surprising, but
- 51:55notable that, you know, companies, some companies have
- 51:58peaked already and that there are a lot of mini hype cycles.
- 52:02I would say, you know, like these companies are making money
- 52:05when they're hypey because they're subscription businesses
- 52:07in in some cases. And so I think that is good for
- 52:10them. But yeah, the, the, the fact
- 52:12that there is sort of mini hype cycles is is notable.
- 52:15I think, I think what's really interesting is that the majority
- 52:19of the revenue we discussed coming from consumer
- 52:21applications, right? And that includes open AI,
- 52:25includes mid journey, includes perplexity.
- 52:28I mean, no maybe. For the record, there are a lot
- 52:30of B2B companies making plenty of money, so this is just where
- 52:33we have more interesting engagement data.
- 52:35But yes but. I do think that a lot of it is
- 52:37with consumer. Yeah, that's, that's my take as
- 52:40well. And I think it, it's sort of
- 52:42interesting, if you think back to the Internet era, like a lot
- 52:46of the companies that kind of emerged first and became the
- 52:50biggest companies were those consumer companies.
- 52:54And then you had this later wave of like SAS companies, right?
- 52:58Like kind of copying a lot of like the consumer Internet
- 53:01mechanics and, and graphical interfaces and sort of adoption
- 53:07patterns. But I think, you know, maybe
- 53:10we're seeing something similar that that consumer kind of leads
- 53:13the way and then and then you get, you know, later, later
- 53:16companies kind of adopting those that technology and B to B.
- 53:21Yeah, I think I, I have a similar perspective to you guys,
- 53:23which is that this seems like the greatest consumer revenue,
- 53:27like boom time I've ever seen. I mean, I just never, I've never
- 53:30seen people sell stuff for money to consumers in in the Internet
- 53:34post Internet era, right? I mean, the whole idea of the
- 53:37Internet was like everything's free, give it away for free,
- 53:39just get as many users as possible, like sell ads later,
- 53:41you know, Google and Facebook and Snapchat and every etcetera,
- 53:44etcetera. But like every single one of
- 53:46these companies is monetizing like.
- 53:51The one thing we haven't been talking about is they're giving
- 53:54people a service that might cost, you know, that that might
- 53:57be more expensive to produce than it is you're.
- 53:59Saying unit economics might be the issue, yeah.
- 54:01Upside down, yeah. Yeah, I think that's fair.
- 54:04But I, I sort of I don't know, I doubt that these things are
- 54:07deeply out underwater on unit economics.
- 54:09I think open AI can afford to sort of incinerate money and and
- 54:12probably anthropic. But to your point, I think mid
- 54:14journey our understanding is a super profitable, you know, my
- 54:17bet would be that a number of these other, you know, companies
- 54:20are are unit economically doing quite well, whether it's like
- 54:23Suno or Perplexity or you know, any of these other ones.
- 54:27So I I think that, yeah, it just feels like monetizing consumer
- 54:31products is like the narrative of the last year and it'll be
- 54:36interesting to see if that continues.
- 54:38All right, that's our show. Thanks for listening.
- 54:40Thanks, Max, for putting all that data together.
- 54:43Sure. Cerebral Valley is November
- 54:4520th. We will have talks from Dario to
- 54:49Anthropic. Yeah, you'll get to hear on our
- 54:52YouTube channel and in this podcast feed from some of the
- 54:54people we're talking about here. And thanks so much for
- 54:57listening. Thank you.
- 54:58Thanks.