EP 136: Zach Weinberg (Co-Founder & CEO, Curie.Bio) on The Tariff Trap, Self-Inflicted Recession Risk, and Open AI’s Fundraise

28:55And I think he's going to let this thing play out much longer than I first anticipated, which is bad. And then you get this self-fulfilling prophecy. This is the part that really matters that I think it's like, even if he unwinds it at a certain point, you've already done real harm. Which is by having an anticipation, I think this is like really important for people to realize, anticipation of higher prices. not even just the prices themselves, but the actual like anticipation of higher prices, which you can see in all the polling data, people are like, oh, fuck, there's gonna be more inflation.

29:30Prices are gonna go up. Tariffs aren't great. Causes people to stop spending. They budget. They budget. They go, ah, this is good. I'm gonna buy stuff now and then I'm gonna stop. Right. And they get more conservative. So they kind of like accumulate and stop. They get more conservative. They don't take risks. They don't invest because uncertainty is like a killer of economic growth. And it's not just uncertainty for businesses, it's uncertainty for consumers. And so when you get consumer spending slowdown, which if you saw, I think it was like FedEx announced, they're already seeing it. Walmart's announcing like soft consumer sentiment, therefore expected spending.

30:05You get a recession. It is a self-inflicted recession. We basically tell people we're gonna have higher prices. Yeah, they're like, oh shit, higher prices. Okay, I'm gonna slow down, which causes, in this case, probably in the beginning, a consumer demand-led recession. And the recession is going to come from lots of different places. But in this case, it looks like consumer demand is going to shrink, which is then going to cause businesses to shrink. And you run that little cycle, right? And like, it's hard to unwind that. The other problem with it, which is like, this is the kind of stuff the global macro people are really scared about, is in a recession like that, where, which seems possible, maybe not likely, but I don't know, 50-50.

30:41We don't have a lot of tools in our toolbox. Like, what are we going to do? Like cutting rate? you know cutting rates i that's going to be challenging it's a it could be inflationary right like is that actually going to be something that's even possible given the state of affairs of like the u.s deficit are we going to be able to like properly bring the tenure down i don't so like all of this that's where the stagflation thing yes uh comes into play which is like uh for people right don't appreciate how bad a uh uh stagnating economy with inflation can be which I've never lived through, but I'm told from economic data when we last had that in the 70s, it's really, really bad.

31:24And by the way, this whole tariff thing, this isn't like there's people on both sides of this, at least at an economist level. Economists, right. If you look at, obviously there's people that argue this is net good, and there's some people that will benefit from it. But no matter any survey, and I was Googling before we came on, I was doing chat TVT, it seems like almost every survey of economists, the academic elite, the, you know, those people, it seems like at the lowest, we kind of get 85 % think it's a really bad thing, bad thing. And at highest, it's up to like 95 with like 5 % uncertain.

32:05Like they did a test, I guess there was an ING economic experts panel that University of Chicago put on in 2018, and it was about steel and aluminum tariffs. The question was, imposing US tariffs on steel and aluminum will improve Americans' welfare. 0 % agreed, 93 % disagreed, and 7 % were uncertain. So it's like, this isn't some thing that I think there's some broad, hey, there's good people arguing either side of this, at least among economists and policymakers. I think this is something really, really important, which I agree with on the stagflation front. By the way, what a horrible brand.

32:47Yeah. Stagflation. Like, what does that even... Part of the reason I think people aren't afraid of it is they don't know what the fuck it means because it's like sticky high prices and no growth. Yes. So it's like... I feel like it went through some like corporate branding exercise that like a lot of people got in a room and they're like well we can't be called like this anymore staxflation that's what it's going to be like i know it's a mouthful but should we just call it like high prices no growth recession because like that's what we're going to get yeah we're fucked uh like nothing good right like and and i i think but you said something really interesting which is like we haven't really had this since whenever 70s 80s you know most of the people who have been through really bad recessions are older.

33:31So like the vast majority of the US has not really been through it recently, or even if they were, they were probably young. And also like human beings, our memories are short. And I think people forget how bad a recession is, and especially one that has like stagflation style characteristics, how bad that is. Like it's not good for everybody. Everyone and we haven't really had one of these in a very long time because for all the stuff, even if you go back to like 2008 where we almost had one from the real estate crash, then the Fed pumped like trillions of dollars into the economy and we had this little dip.

34:10But if you zoom out in a stock chart, it's like hard to see it. And so I just think we haven't had real problems in a while and so we forget why. And it's like the vaccine shit. it's like people are like, yeah, but we don't really get polio. So what's the point of the vaccines? And you're like, well, that's why we don't get polio. Exactly. You're so close to getting it. Yeah. They're like, well, we don't really get recession. So like, we're just going to tear up the shit out of people. You're like, you know why we don't get recessions? Cause we have this like giant cheap shit economy, which is really good for everybody.

34:43And so I guess we just have to like relearn the lessons that, you know, once they get like a little stale in your mind, you forget. get uh i'm trying to give you been i got out of school in oh eight so like right into that crash yeah i was i was 10 so i i like did my internships into that crash but by the time i i came out it was sort of boomerang back around see isn't that amazing two years that's a tight almost non-existent you know and a bunch of people lost some money in like residential real estate that's that we really contained it. I worry this is going to be very hard to contain because it impacts so many things.

35:30This is kind of a transition, but also at some point, I guess the next time we do this, you and I were both pitching a guest or two that I think would be good to talk about criticizing the left and some of their policies and problems that have led to, I guess, maybe a lack of economic growth or at least making things difficult for people to move forward and actually innovate in a meaningful way. So I guess at some point we'll do that. One thing that's underlies a lot of this, and it's interesting because, you know, in venture, and I'm curious if you guys talk about this as a firm, but the difference between like a fixed mindset and a growth mindset.

36:13And I know that that's like very, you know, in, in vogue or whatever. Maybe explain for people that, uh, that don't know the difference. Well, it's used in multiple contacts. I mean, I think people in the venture world have used it in different ways, but I, you know, where I've seen it is really about like individuals, you know, like, are people kind of like stuck in what they're capable of now, or can you actually like turn them into something better. And so like people have growth opportunities, they can get better. From an economic standpoint, I think of it as like, if you view the economy as a fixed pie versus a pie that can get bigger, that's really the big one.

36:51That's like the fixed first growth. And the history of like the United States economy is the pie got bigger. And as the pie gets bigger, everybody wins. Now we can debate redistribution within that pie. And that's like tax policy and, you know, unemployment and how we like move the money around. But the pie getting bigger is better for everybody because then we have more money to redistribute. And Clinton got this, right? If you go back and you listen to the Bill Clinton speech, like the true, I think of Clinton as like the last true Democrat. And like, if you listen to some of his stump speeches, even Hillary a little bit, mostly Bill, he is talking about growth.

37:34like the democratic party used to be a growth party and you know they would argue like well how are we going to grow faster like these are the best arguments for america it was all about like grow faster because everybody recognized that it's better to try and redistribute two trillion dollars than one trillion dollars right like a five-year-old understands this do you want the 16-inch pizza or the eight-inch pizza and they're like well i could probably the bigger one and so we've lost sight i think i think the democrats in particular have lost sight of if you grow the pie all the like equity and redistribution stuff that they care a lot about just gets easier because there's just more money and we somehow we've lost sight of that and their messaging has been anti-business very anti-billionaire which is related to anti-business and it's not like i care so much about billionaires but it's it's the idea of like the democratic party has gotten to this fixed pie mindset and it's really bad and if you go back by the way on your clinton point like if you go back and and read the uh internet bubble and like a lot of how those things came to be there was pretty close uh alliance between silicon valley and like the Clinton-Gore administration.

38:53I know people like the meme Al Gore inventing the internet, but he did sponsor the legislation that funded the early infrastructure of the internet. And the administration itself was pro-innovation, pro-entrepreneurship, light touch on regulation. It was a lot of things that were basically catnip for Silicon Valley. Bill Clinton signed the telecommunications act in 96, which was the first major overhaul of, of us telecom law in 60 years. And that led to a lot of groundwork for the internet explosion. And so like, we're not that far removed from this, um, from this time that everyone was sort of like pushing this agenda forward.

39:36I actually saw, um, an interesting, uh, there was an interesting chart about like the coalition and the profile of the Clinton voters versus today's voters. And there's there's it's interesting how much that's like federated out that the educated high income people, the top of the top are mostly still pretty liberal. But the sort of, you know, upper middle class like that group has really federated from the coalition that was like, oh, maybe it's Bill Clinton, maybe it's George Bush. And they flip back and forth to now there's there's not really a party I think that uniquely resonates with that group anymore.

40:15I feel like it's a lot of the suburban dads. Like they're not pro tariffs and they're not, you know, a lot of the woke stuff or whatever, like a bunch of the things that they don't appeal to either of those. And so there's this there's this party that maybe once was a Bush voter or a Clinton voter that is no longer, you know, has it has a home to the same extent. I mean, yeah, 100 percent. there's like a weird disenfranchised set of people that tend to be like, kind of like upper middle class educated. I think it's happening now. You and I were talking about this at one point. And I can only speak to like, whatever, techies, humans.

40:54But I feel like there was a deal that you and I signed up for back in like, whatever, 2008, 2010, we're getting out of school, which was basically like, here's the deal with the Democrats and why we will vote for the devs. we like socially liberal policies. We believe in things like gay marriage and kind of, I kind of think of it as like, do you? Like, I don't think the state should really tell us what to do. You do you. We're going to support those liberal causes because we think they're like good for human beings and free trade and all this kind of stuff. And we'll pay higher taxes. Like, we'll eat it.

41:33We'll eat it. Like, I'm happy. Oh, whatever. 4 % more. Like, who fucking cares? Right. And I'm willing to do that. Because the rest of the Democratic Party was like pro-business, pro-growth. And so that's the trade. We like agreed, you know, it wasn't like an explicit agreement, but that was essentially how I felt. And then sometime between then and now, the Democratic Party was basically like, business is bad. And it's not all Democratic Party, but it's a lot of rhetoric that I think got lost, especially in state level of like, business is bad, criminals aren't bad. It just like went off the rails.

42:12And that's why I think you saw a lot of Silicon Valley types kind of like move to Trump end of last year. Let's say they were like, you know what? Fuck this. I'm done with the Democratic Party. They think I'm a bad person. And I think a lot of tech people and finance people too, not that we matter in the grand scheme of things, but like felt that. I don't know if you felt that. Yeah, I mean, I think the two moments I reflect on, and I'm sure we could actually have someone on to talk about this, but the Occupy Wall Street kind of 2011 period of time, and then kind of the Cambridge Analytic, like 2016, Facebook, like all of that, the distrust of social media, which now both parties, I think, equally kind of feel.

42:59Um, but I feel like those are kind of the, the big points in time. Uh, Bernie obviously was always kind of a member of the, the, um, the Democrat or whatever, uh, whatever he calls himself, but like broad umbrella, he, he, uh, he's a broad part of the democratic party. And, uh, I think that plus like Elizabeth Warren and, you know, that whole period of time, I think that kind of pulled it in that direction a little bit, which I, I actually, if you reflect on it, I think from a first principle standpoint, 20 years ago or 15 years ago or whatever it was, I think you and I, we were probably leading with social issues first.

43:40And it was kind of like, oh, you know, the economy. I think it's really important. We led with social issues first because we felt like there was this fundamental agreement that business, pursuit of profit, capitalism, free trade, like let human beings grow was there. We didn't have to talk about that because I just, you should go look at the democratic rhetoric from back then. There's just growth, growth, growth, growth. And so, okay, we had, I think of it as like a luxury. Like we had this luxury that we could talk about social issues because the underlying economic stuff was clean and stable.

44:22And that social trade is unwinding, right? Where it's kind of like, well, yeah, obviously, like, I support things like gay marriage and what, but you guys are coming after, like, the basics here, which is, like, allowing people to pursue a living and make money and all that. And now I'm kind of like, I'm not sure I'm willing to make this trade anymore. You know, the Democrats are like, profit is bad. like what are you fucking talking about like making money in america is a good thing it is a good thing like why can't we just acknowledge that profit is a good thing it means you made something that people want to buy like profit you just like pull profit out of thin air you know like it has to come from like making something that people want to purchase you hire people it's sustainable you took some risk you by the way you take those profits and guess what you do with them you reinvest them into u.s businesses and startups and bad venture funds like yours and mine you know what i mean like like the money cycles back through not as tax revenue but as investment and and i think that trade with the dems kind of like unwound really really poorly people like oh look at this trump guy he's like a businessman you know like in theory uh let's get on that train how bad could it be i really felt that i felt that with a lot of my friends where i was like you are kind of a trump supporter now aren't you and they're like well the other one's really bad and this guy seems like he's like pro business let's try this and that that that little shift seems to have unwound very fast because now even the most conservative like former damn currently conservative friends i have and like i don't know they're mostly like whatsapp groups are kind of like yeah this is bad this is bad yeah well i think that's an interesting thing and then we can put a pin in sort of the the macro politics stuff but one of the it felt like the first term of the trump administration there was kind of this uh in broad strokes, like there was the Jared Kushner sort of side that still had a global macroeconomic mindset and was moderate, it seemed, in a lot of ways, as like a mitigating force for some of the worst inclinations of populism, it seems at least.

46:37And I haven't discussed this with anyone to know behind the scenes, but at least that's my outside-in observation. And then there was kind of this populist, nationalist thread that felt probably more core to the MAGA movement in a lot of ways. One of those, the Kusher side was kind of like the broad strokes, like father-in-law bucket of the conservative party, not mine specifically, but general. And then the other side was more of this nationalistic populistic movement. And it felt going into 2024, you couldn't really tell which side of those it was going to break to. It seemed like the rhetoric was populist, but we had seen the proof in the pudding.

47:20Everyone sort of said going into it, take him seriously, but not literally in all the stuff that he was talking about. And it's clear we're way on the populist side right now, which I would never have expected. I thought the stock market would be the global voting mechanism or the national voting mechanism by which we would kind of keep score and then ultimately decide if things were working or not. And it's clear that's just not the case. Similar to you, I thought this stuff would have been unwound weeks ago. So I don't think it could have gone poorer. Like there was some stuff they were doing I was actually really supportive of.

47:57And, you know, we were talking about regulatory reform, unleashing the U.S. economy. We were talking about unleashing our energy resources, both oil and gas and modern solar and nuclear and all these things. And then we just started to do all the stupidest things humanly possible in two months. One of the things I'm most surprised about is the Doge Elon Musk thing. Because this guy is maybe the single best entrepreneur of our generation. And he's got, he's top three, you know, in the sense that he has started multiple massive companies. Like, look, someone starts one awesome company. It's impressive.

48:45Starting three awesome companies. Like, I don't think people appreciate how fucking hard that is. That is so hard. And to do it in different industries, in hardware and two of them, like this guy is, yes, he didn't start them all, but he's pretty amazing. And so you assume that some basic ability to like learn something new and go in and come up with a plan and talk about government efficiency. That seems like a good idea. Like I was, I'm like, oh my God, Elon Musk is going to be in this administration talking about government efficiency. Like this is great. Like we need this. And the execution, not just the rhetoric, but the actual execution of it has just been, it's like watching a 20 year old make a bunch of like operational mistakes.

49:31They remind me of like an inexperienced founder. And I don't understand like how and why that happened. Like why not just spend two months like doing homework like you would normally do in a startup? Like figure out how this thing works and understand where it's inefficient and then come up with a plan and run it by experts and do all the normal things you're supposed to do in a startup. And he did none of them. Well, I think we sort of saw, I mean, this is clearly a replication of a playbook that was run with Twitter in that way, where I think it was far more, you know, measure once, cut three, four, five, ten times rather than the other way around.

50:13And so I agree with you in principle, but I think there is the precedent of where it's made sense. Yeah, it's going to be, I think Doge could be its own, I mean, obviously, very long podcast for us. So maybe let's pause there. I do want to transition. I mean, there's been a lot of open AI stuff this week. And I guess maybe that's just a universal truism at this point. Like we could probably just say any week there's been a lot of open AI stuff. So I guess to rattle them off, and I'm curious what you find most interesting in this. So they closed a$40 billion round of financing,$30 billion from SoftBank,$10 billion from the syndicate of investors.

50:52I think the initial funding, this is interesting, will be about$10 billion, followed by$30 billion at the end of 2025. And SoftBank, I guess, disclosed that maybe its total investment could be slashed as low as 20 billion of OpenAI doesn't restructure to a for-profit entity by December 31. Also, I think we've seen OpenAI absolutely explode. So there was on Monday chat, GPT now has 500 million weekly users up from 400 million last month. So added 100 million in the last couple of weeks here. They also released their image model. They also, or a new version of their image model. They also said they were going to release open weight models as well, which I think is something that people have been calling on for a long time.

51:44And I think Sam indicated was on its way. I'm curious, anything of this that jumps out to you or that you find most interesting? I mean, a few things, I think only because I'm in, you know, economic mode, just from our purpose, like, can we just pause for a second and acknowledge how amazing it is to have like five chillion companies startup and big like all competing at the same time like the product innovation the speed of it the cost of it and by the way like they're almost all american companies it's really incredible how much of your activity that would have previously been reserved for google search do you think has now converted over to chat gpt or perplexity or gemini or whatever probably like a solid 10 20 percent 10 20 percent of your act you're still doing 80 90 percent on google i think of it actually more as like going to like fix versus growth i actually think i do more research now like i think of it as like there's a like the pie has grown for me there are things that i now go chase down using either like perplexity or chat gpt that i might have just like not done before um and then some subset you know shifts over from google i also like to test the google i'm in the beta so like the native google ai mode that they just launched so is that random or did you pull some strings to get into uh into it i just like signed up for the thing and whatnot this wasn't like my company to you 15 years ago uh can you give me uh can you give me some preferred access no i haven't pulled that card uh ever probably it probably doesn't work anymore.

53:20I think that's probably, it's probably worn a little thin. After 15 years, there's a statute of limitations on it. I would say my thing is converted way more than that. I haven't actually tracked it, but my guess is 60, 70, 80 % of stuff. And it is growing the pie as well. It's amazing how many things I now default to going there. I had some stain on my kitchen counter and I didn't know how to get it off. And so I took a picture of it and I didn't know what it was. And I uploaded it to the chat GPT and I said, Hey, what do you think this is? And they gave me like a diagnosis of it. And I was like, all right, well, what are the techniques to get rid of it?

54:01And these multimodal things that previously I would have spent so much time like trying to debug it and testing or even like there's so many little processes that I think I previously would have spent time on YouTube searches or just clicking around to different forums. And now it just lays it out in such a clear synthesis, my behavior has totally shifted. And so it's just, it's remarkable how sticky that's become for me. I think like, I would start, I'm going to make a little bit of a bare case for open AI, but I always want to start with like, this is just really good for America. Like Apple, Amazon, Google, Anthropic, open AI, perplexity, some of these voice models, like 95 % of the innovation is here.

54:51And at some point, we should like take a step back and be like, this is pretty awesome. Let's not fuck this up for the next thing. Leaving that aside for one second, OpenAI specifically and why I don't, I wouldn't be investing at this price. And I was even skeptical at the able. I think there's like three bear cases that you could argue. And let's see if any of these resonate. my favorite new one is soft bank tends to top take everything. So there's just that like soft bank has a history of, you know, like investing at the peak and losing all their money. So there is a signaling association with, with, with that.

55:31It's not universally true. There's, they made some good bets in history, but yes, in recent history, I would say there is a top tick element. You know, they have those two phrases. The broken clock is right twice a day. And what is it? Blind squirrel finds a nut. Masa seems to be an incredibly blind squirrel, But like, okay, we'll leave that aside. I think there's kind of two bare arguments that I would consider. One is just the margin profile. I think we've talked about this before, but like this isn't SaaS style margins in these like foundational models because of the CapEx, because of the amount of infrastructure that you have to build to keep pace.

56:10And you're competing from like a CapEx spending, data centers, inference models, all this stuff with some of the largest balance sheets in the world, right? It's Google and others. And so what looks like incredible revenue growth, and it is incredible revenue growth, it's like the fastest I've ever seen, but it's not SaaS margin revenue growth. And so the question ultimately in any business over the long run is like, can you make money? Not just like, can you grow revenue, but can you be profitable? And then, you know, applying a multiple to that. And so I think there's some very fundamental questions about can open AI as a business be profitable?

56:49Not can it grow, but can it be profitable? Now, they're doing all the right things. I think building this consumer brand and subscription business is really smart because it makes you sticky. Which leads to my other question, and maybe they were just early in the cycle, because I like your stain example. that functionality which is basically multimodal like I can text and image use text input and use image input in real time that is going to be native in your phone within two or three years that is not going to be like open up the app at least in Android because Google's ahead that is going to be Android native functionality like all the Gemini stuff because you can do that same thing in the Gemini app It's not just ChatGPT.

57:38Maybe it's just not as good. That stuff's going to get built into Android natively. And so are you really going to go to the ChatGPT app when the functionality is as easy as Google Search? Have you seen that little Google bar that they stick at the bottom today, which has image input, voice input, and normal search? You're going to have multimodal input. Where I think OpenAI has this really interesting advantage right now is that Apple is so slow. And so there is no native iOS functionality that ChatGPT is competing with. And so in that world, I do think they have an interesting runway. The question is just how fast can Apple and iOS catch up to move this functionality from a standalone app into native OS?

58:26That's what I would be concerned about. I would be concerned about being front-run by the phone companies, really. There is a problem of defaults. And I guess in some ways we can go back to either the browsers in the Microsoft generation or the browsers on the different phones, the search on the different phones, the maps on the different phones and all of that. I do think that there's power of default. And I think we should always just assume consumers are lazy and they're just going to do what comes easily to them or what they recall easily enough. I think we probably would agree that the B2B side of this market is going to end up federated in the way that a lot of B2B markets, probably not too dissimilar to cloud computing, where, you know, you have an AWS and GCP and an Azure.

59:20And, you know, and so maybe that's you have OpenAI and you have Azure again, and maybe you have Anthropic or whatever it is. You have some version of these folks and maybe with the open models, it just ends up being the cloud hosting providers, or maybe it's OpenAI and Anthropic and whoever else, XAI that can do that. But I guess let's put the B2B thing to a side because I think that's sort of different. Then on the consumer side, I do think the power of defaults is a risky thing for this business, for sure. That said, they're at 500 million weekly users. And I have just had the power of delight or the amount of delight I've had from using OpenAI.

1:00:01There's been, I would say, so deep research was like, oh, wow, this is so cool product for me. And the native image generation thing more recently was like, oh my gosh, this is just so cool. And so I do feel myself increasingly sticky, even though I had Gemini on my phone, I have Grok on my phone, I have perplexity on my phone. I am increasingly getting more and more drawn to using OpenAI just as my default every time there's one of those breakthrough experiences. And then at some point, just based on the user, that's because they are ahead. Like we should acknowledge they are ahead. But yes. For how long and at what price point?

1:00:46Yes. And so it's a reasonable question of what is like, do you think is it is there a tolerance? What is a switching cost associated with it? Right. As I understand, there's a lot of good products that have been built that ultimately, once I got going on using the existing. product. I just never thought about switching to the other one because it's better. And there's inherent defaults associated with this. I obviously ostracize you for having green messages on your phone. And so there's a lot of little friction points. And I just think the data, the corpus of data and the amount of feedback loops and the behavior that can be personalized to you when there's more longer context and memory associated with the product itself, I just think they can continue to build on this viral loop of it.

1:01:37And so I worry of the two things you mentioned, I worry a little bit more about the CapEx side of things and just the cost of continuing to push forward versus the defaults side of things. But we'll see. I mean, obviously this is being played out and you're dancing with elephants in a meaningful way. Well, I think this question of how long can they maintain a true product lead is the fundamental question. Because I agree there are defaults and the longer they're ahead of everybody else, the more those consumer defaults will lock in, even if Apple catches up. And that's a huge advantage. Now, it's not free to run ChatGPT, right?

1:02:16There's a lot of cost in trading the models and then on the inference side as well. So it is expensive to pursue this and to run it. So they have to charge for it at some point. And my concern, and let's just use the Google example, would be if Google can catch up, and I'm going to leave that as like a giant execution if, right? If like Gemini can catch up to chat GPC in quality, and they're kind of like getting there on the image generation side, and maybe they can get out of their own way. Google can run Gemini at a loss because they can monetize it with the ad machine. and that is a dynamic that open ad cannot do like eventually like you have to make money and if google's just going to use its giant balance sheet and its ad machine so you run you know they could run this thing they could run gemini at a multi-billion dollar which was direct cost loss for a long time decades and just monetize on the back end eventually with ads open a like the price point that's kind of the thing i would be really concerned about with open ai is like just erosion of the price point by google and maybe eventually apple the fact that apple is so far behind is really interesting like they are in trouble and they're going to have to buy something and we'll see if the government allows them to but it would not shock me if at some point Apple goes $100 billion for Anthropic.

1:03:50Like, they might need to. The thing with Anthropic is the consumer business just isn't there today. Totally, but if you take their technical infrastructure and then you layer it into the mass distribution on the consumer side that Apple has, you can think of OpenAI as Anthropic being like the deep mind. of Google, right? It kind of like launches their infrastructure. I mean, that would be the most logical outcome if I'm Apple is like building this in-house is going to be difficult. And every day we're not there or like losing ground in the way Microsoft lost ground on the phone. So I don't know, but margins, price point.

1:04:36If we look back on like corporate battles, this is at least in terms of like US, which I have far more familiarity with. I mean, we can, obviously there was Apple, Microsoft in the 80s to 2000s that was particularly, you know, heated. And then to some extent, Apple and Google, Google and Facebook and more recent history, Amazon, Walmart. And then, I mean, we can go back and come up with a bunch of ones way back, Coke, Pepsi or whatever, Ford, GM. But this is one of the most interesting chess boards because it's not just there's a lot of global supply chain considerations. There's incumbents and disruptors.

1:05:25It's expensive. It's very capital intensive to do it where Google was able to operate in a pretty light CapEx way for a long time without burning a ton of cash. And so this is just like, in terms of a business case study of how this is going to play out, it's the most fascinating, at least that I can come up with in recent history, for sure. I'll layer in maybe one last thing I think is interesting. And if I were opening, I'd be thinking about this pretty deeply. Consumers are very sensitive to price. and like the fact that like Netflix has this like incredibly massive free tier blows my mind it's like hundreds of, I mean it's insane how big it is and Netflix is like 12 bucks a month and what's chat GPT is like open 20 a month or something like that yeah it's a little complicated to compete in the long run with free and that's where Gemini is going it will be free.

1:06:29And there will be paid versions of it, but the native stuff is going to be free. That is a hard thing to compete with. So I'd be really worried about it. I ultimately think OpenAI will have to. I know Sam's resisted or talked about resisting this, but I just have to think that they will need an advertising supported model. I know he doesn't want to do it, but it just unlocks such a huge part of the population. There is this hedonic or insatiable thing that people are willing to readjust to the new normal of what their expectations of free would look like. And if you had shown me what the image generation could have done five years ago, I would have been like, oh my gosh, I'll pay$20 ,000 for this.

1:07:13And now I'm like, this is bullshit. You're going to charge me like$20 a month for infinite ability to change anything. The way I would go about this, if I were in charge of products at OpenAI is I would be thinking about non-monetary lock-ins. Because if I'm going to compete with free, I better have some lock-in value that isn't just essentially a high-quality search query. And I think it has to be, I'm going to kill myself for saying this out loud, but I think it has to be social networking. I think you need some reason that it makes it hard for you to leave OpenAI, ChatGPT, even if something else is cheap or free.

1:07:54And if the tool is utilitarian, which it kind of is today, find something, make an image, do a query, whatever, like I have a thing I want to get done, that's switchable. Because if you could just do it cheaper over here, like people will switch. If I have like friends and social and sharing and something native, I actually think I would go after like messaging within chat. I think people would message and the way they kind of do inside Bloomberg chat. And I think you could backdoor an interesting messaging platform there. Then the threat of free from Google and Apple, I'm not as concerned about because there's real lock-in.

1:08:32But I don't know. I assume they're thinking about this. I hadn't thought this through. Yeah, I mean, they have smart, they have WILE and a bunch of interesting smart product people over there. So I'm sure they're thinking about all these things. Yeah, it's interesting. Like Spotify versus Apple is an interesting analogy that I would have thought my default behavior just conceptually from the outside end, I would have said, okay, I started with Apple Music, but I've now moved to Spotify. But when they go back to a latest thing with Apple, I'll go back to Apple. That would have been my like zoom out thing.

1:09:07And I don't know about you, but that just hasn't happened for me. I've stayed with the default of Spotify. I haven't used Apple enough to know the nuances, the product difference of it. But I've never even felt tempted, even when they were doing the exclusive album releases and trying to find all these different things. And maybe it's just a shittiness of the product thing. And so to your point, like if you can compete on price and be free or if you can compete on product in some meaningful way, you at least need to get to a parody level. But just an analogy that I've thought about in the past of like that didn't change my behavior in a meaningful way.

1:09:43I do think it's a price point materially different that we're having a totally different conversation. Right. And it's not. It's basically the same cost for Apple Music versus Spotify. The other thing I'd also highlight, too, is Spotify has really good... It's got improving margins. Apple doesn't have an ability to monetize without charging. So I agree it's an interesting native versus app comparison. I think the economics of these businesses... I will say what Spotify has done against the music oligopoly, I never thought... I mean, there's just a couple of labels that theoretically should, they colluded to keep the prices where they were in the industry.

1:10:22And Spotify has just continued to generate the user demand side of things that have given them more and more leverage. And then AI discovery, now they're the ones with their algorithms that can make or break artists. And so now they have power that the labels no longer have. So it's a very interesting case study. There's probably a lot of lessons there to be learned for open AI. I'm curious, before we hop here, what's your perspective on the wild card and all of this? In the last week, there's been the combination of the artists formerly known as Twitter and XAI. And obviously, the bull case in this is that's such a unique corpus of data that Twitter or X provides.

1:11:04I'm never going to convert the saying X, by the way, but that it provides. provides. And so theoretically, there should be the most proprietary data set to learn things on with Brock. The flip side of that is like, I don't know where that business is right now compared to Anthropic or OpenAI or some of these other names that we've talked about. My guess is materially, materially behind. Yeah, I mean, it's another competitor with a big balance sheet and a willingness to look at super long horizons and a willingness to lose money for a while um they'll move fast and you know they're not going to have any of the breaks that like a google will have to have internally so it's another interesting real threat i mean i actually think like grok basically when they first came out removing all the or most of the restrictions on image generation was like the key to i still use it for image generation now open ai seems to have rolled back their by their ability for me to mess around and make memes and stuff uh for the content policy restrictions related to copyright i can't make you south park or you uh you know whatever ghibli style anymore like i like i could two days ago uh so it's like can you make can you make a black nazi remember that was the original one yeah they're like in in gemini it was like the only nazi you could make was like you know because it had to be equitable and then you know the the fear of like i don't know if we can have these images and i think grok was like i don't know make whatever you want and that's really what people that's where this is clearly going and besides legal copyright issues which will always be there for everybody.

1:12:56It seems one of the benefits of Grok is just like, we're going to remove all the guardrails here that are kind of stupid. Use the technology. Just free speech. Come on. Let people do this stuff. As long as they're not breaking the law or making child pornography and stealing copyright content, which will get litigated anyway. But yeah, I don't know. That's another entrant. And a big one. and maybe that's who Apple picks up at some point because they can solve both their like social networking thing and the model all in one company. I think Apple has nothing to do with Twitter. I'd be very far away from that one if I were them.

1:13:41But yeah, yeah. I mean, let's see.

1:14:00We'll see you next time.