Ep14. Public Market Volatility, AI Air Pocket, $GOOG Ruling | BG2 w/ Bill Gurley & Brad Gerstner

27:11It actually causes the market to go down. You know, our team was looking at some data in 2019. In 2018, there was fear about the economy slowing. The Fed starts cutting rates at the beginning of 2019. And the market was down big, right? It's down big because it thought we were heading into a recession. So it took a, you know, probably six months for the market to rebound from that. And a much worse version of this happened in 2001 that you and I remember well, right? The Fed continued to raise rates in May of 2000, worried about the dot com boom and that the market was overheating, but they were way behind the curve.

27:51And we all saw the economy really slowing down and signs of the dot com bubble bursting. They were late to the game. By the time they started cutting rates in the beginning of 2001, it was clear they were behind that the economy was cascading into a recession. And the Nasdaq was down over 20%. And so I don't think, you know, as market participants, when we're weighing all the risk reward here, we're not sitting here saying, oh, the Fed's going to cut rates and that in and of itself, you know, fends off a recession. We need to make sure that we continue to see earnings come in strong. You know, that one, that particular situation, there's one thing that people seem to have a really hard time remembering after the famous quote about irrational exuberance from Alan Greenspan, the Clinton administration actually lowered capital gains tax rates, which fed into that problem.

28:45And probably shouldn't have should have been raising them. Anyway, so, so let me ask you this question. Once again, macro bill here. What are the chances that you get probably the worst situation of all where the Fed cuts and nothing positive happens? Like there's no positive impact to growth. Is that a possibility in this case? Yeah, I mean, listen, I think the markets are already doing the Fed's job, right? Remember that a lot of variable rate debt is priced off of where the tenure is and the tenure has already, you know, come down. So people's mortgage, you know, you can already go get a cheaper mortgage.

29:28You can, your credit card debt is already coming down, et cetera. So this is the dynamic. And the marginal impact of the Fed lowering to those long term rates is not going to be that big. Correct. I don't expect it to be that big, but it does signal directionally, right? That the market is right. Remember, the market will ultimately adjust to where where the Fed is. So the Fed, it's not to say that the Fed funds rate is, you know, is not important. I would say this, what's far more important about whether we end the year higher or lower is not whether the Fed cuts rates. It's what is the economic growth?

30:06What is the unemployment? What is the CPI? What's really going on in the underlying economy? And again, just again, telescoping out here. So you have this concern about taxes. We're not going to know the answer to that until after the election. You got this concern about recession. You know, the Fed can cut rates in September. That doesn't make that concern go away. That concern is still going to be out there until we see additional inputs on economic growth, you know, come in. And then of course, this third one that, you know, we were talking about Bill, which is just concerns over this AI air pocket, which you've been beating the drum on.

30:45Okay. Yeah. Well, I mean, look, I think to a certain extent, that's played out, right?

30:54I would have called AI stocks in video being the most obvious, but I think the max seven, a lot of them are exposed to this. They've all come in, right? And so you and I had the discussion, you know, we'd been having it all summer, but then the Goldman thing came out and the SCOA thing came out and then the stocks retreated. So could you make the argument that that bubble's been popped already? Well, I, you know, again, like I said at the beginning, I think these are healthy consolidations, right? Like what you don't want, what a bubble is made of is when everybody believes when there's no talk about an air pocket and when stocks just, you know, elevate to ridiculous multiples, I think here, you know, you've seen Nvidia come down from 140 to I think at bottom, dead a hundred or 95 bucks, it's probably a 105 today.

31:46And so, but we did learn some things in the earnings reports of the last two weeks, Bill, that I think we should call out in terms of what meta said, what, what Google said, what, what, what Microsoft said about their CapEx and about AI spending. Here's some charge from my partner, Frieda Dwan. And if you look at the big four, they're going to spend 220 billion this year on CapEx, you know, and if you look at their statements on these calls, right? Soon, Dar said it's more dangerous to underspend than to overspend, right? Sacha said, they're still capacity constrained. So they need to continue to make these investments.

32:27So I think in the short run, all of these companies meta said basically the same, Amazon the same, they're going to keep the hammer down on CapEx because they're seeing the return, their capacity constraint, and they're worried about not making the investments, right? So if you compare this to the quarterly and the annual expectations, we took people through on the pod for Nvidia, right? Here's this chart. We talked about this $2 billion that's got to get, or $2 trillion that Jensen said would likely be spent between now and 2028 on data center build out. And, you know, the green line descending line in the middle, right?

33:09This is if you just take the consensus numbers for Nvidia and what that means for their share of the market over this period of time. So the consensus numbers, while Nvidia's numbers continue to go up a lot, right? And they're going to report in two weeks here. But if we look at Nvidia's quarterly expectations or their annual expectations, so this year, they're expected to do about $136 billion of data center revenues next year, that goes to $166 billion. So up by about, you know, $30 billion, obviously the rate of growth is slowing a lot. But when you compare that to the statements out of the big four, you know, it certainly seems to foot, right?

33:52It seems that they're going to continue to spend. And I think the real question as to the air pocket now, right? Like this could have been very different. They could have got on their earnings call bill and they could have said we're pulling back on CapEx because we're not seeing the return. Okay. None of them said that. They all said the opposite. But I think the question that remains out there is whether they're going to see the revenue on the other side of the equation, right, to justify that return. And they've all cautioned us that they expect that there's, there may be a lag between these two things.

34:28You know, what was interesting, the CFOs of Microsoft and Meta, Amy and Susan were quick to point out that if the revenues don't materialize, that they don't actually have to build out the data center shells, right, that they're acquiring. They said these are 15 year long live assets. And so we have flexibility in how, you know, we manage those assets. Yeah. And obviously Meta has other uses for internal use and could redeploy those things. And cool. Well, I have three thoughts about this. You know, and one of them does go back 30 years to when I was a Southside analyst and I followed the PC industry through really through a Tadek, like the big growth years of the PC industry and PC demand and growth was overestimated and underestimated five different times, you know, where where where you would have these cycles.

35:27And I was right. It was interesting because a lot of the commodities would become in short supply. So DRAM was one and hard drives hard drives were this really interesting wild market where you would have boom bus cycles and shortages and there were always, they're almost always out of step with reality. And I remember there's just this bizarre thing where you had to buy hard drive companies when their PE was the highest because that was actually when everyone had the worst view possible of their earnings possibilities. Right. Right. And then you literally sold it when when the the PE was the lowest because the net income was was at a peak.

36:14Anyway, one of the reasons that happens. One of the reasons why that's so out of whack is that bringing capacity on to build a hard drive or a DRAM is slow. It's not an overnight process. And the one thing I would say about these data centers and I was I was just kind of my brain was triggered on that from something I read on Twitter this morning where someone was literally walking through the steps you have to go to do a fire land to get electricians on board to do all these things. And whenever you have this kind of really long cycle time to actually bring up incremental capacity, it's easier to miss supply and demand.

36:57And so anyway, I think part of what's causing me to have a bit of anxiety is that fact that and I and with and I also think when people are talking about Nvidia, they talk about data center growth as demand. But because everyone the end users buying these things as a cloud service, I kind of like, well, that's supply. That's not demand. Like people are aggressively growing supply. So I don't know. It's super interesting. I think from a venture point of view, you're hopeful that there's so much demand for this stuff that they never catch up. Exactly. When I look at it from a from just from a a neutral observer standpoint, though, I do worry that you could get out over your skis.

37:45And then two other things I would add to that. This is one of the very first waves where the incumbents were eyes wide open at the beginning of the technology shift. If you go back and read the innovators dilemma, you know, the the reasons startup companies are able to come in and take shares often because precisely because the incumbents are slow -footed and aren't there. Here, they moved quick. I think Microsoft and Sacha are just the poster trial for this. Like here is a company that's very what 35, 40 years old. And everyone thought, you know, had had had had its better better days behind it.

38:27And they came in and moved fast and people loved it. And the stock was rewarded for it. And I even think their earnings have been positively impacted by it. And so that is so they moved big. They moved fast and it worked. And so now I think that's in their psyche, right? And so that means they're going to keep playing that game because it's working. So that's that's another thing. And then lastly, and this is pure conjecture on my part. So I did not hear this from anybody. Everyone can just say Gerlis out on a limb here. But my reading the tea leaves, and this relates to the point I just made, is that they don't want the independent AI companies to take, you know, what they see is rightfully theirs.

39:19And we've been through a couple of waves here where we're capital availability's dramatically improved for private companies. And you know, you and I lived through this with Uber and and Division Fund and Mossin, all the money and and DoorDash got all the and it really perverted the market. And so now I think incumbents are aware that if companies get breakout potential, they can raise unlimited capital. And so to a certain extent, I think they're pushing back. They're saying, look, this is this is a game we're in. You look at Zuckerberg like, you know, with the open source stuff, he's just being very aggressive.

40:01And I think one of the objective functions may be just precisely to make it more difficult for open AI or inthropic to raise, you know, $10 billion that they need to fund the next two model builds. And now you've got people, you know, trying to project income statements and balance sheets of these companies and they're speculating about whether they might need to raise again or not. So I think all that's super interesting and unprecedented. Very insightful. What's what's fair to say at this point, you know, again, in the context of the risk -facing public markets, we were worried about the AI air pocket.

40:43It seems we've traversed the 2025 CapEx concern cycle, right? They are going to continue to spend. What we don't have insights into is that end user demand that you talk about. And I think it's really interesting why you say the bigs are going to continue to spend. And in fact, some of these take -unders, you know, the characters, the inflections, etc. certainly suggest that competing against the bigs. I mean, in the case of open AI and inthropic, right, they partnered with them. They're owned a lot by, you know, Microsoft and Amazon at this point in time. So it's hard to find somebody who you think can go the distance at that scale that's truly independent.

41:28And by the way, regardless of whether the investors got paid, maybe we can go deeper in that later on these take -under deals. It is a capitulation, no matter what. Like even if, even if it were a normal M &A deal, and it was at $3 billion and still a great return for investors, the founders are saying in the inflection case in character AI, they're saying, hmm, not sure I want to keep playing the independent game anymore. Yeah, it may have just been very difficult to raise private capital in an up -around, you know, in this environment. I mean, because you know you're going to have to go compete with Google, with Microsoft, with Meta, you know, they're not going to stop spending.

42:10And they have a printing press that's spitting out $10 billion bills in the back room. Well, and I've got hundreds of billions sitting in cash anyway. Exactly. So yeah, they like, and this, I think you brought this up, or maybe we saw sides to talk about it, like, okay, we're spending 50 billion a year on CapEx, but look at all this cash we got sitting around. And the FTC says we can't do acquisitions. What are we supposed to do? Like, then that's another factor. Maybe that's a fourth factor. Yeah, three things you can do, Bill. You can buy your own stock, you can issue an dividend, or you can buy Nvidia chips.

42:47They're doing it. Yeah, exactly. I think they're doing all three. Yeah, they're be interesting. You know, I, they're obvious, huge wins in AI. You know, we've already talked about Tesla and full self -driving. I think all the core AI opportunities, non -Ela Lim, where they've actually figured out how to go from A to B and get better at what you're doing through AI. Those are fantastic. I think the the language oriented things around customer service, and where you're literally replacing the notion of someone either talking or doing a search. I know you're a big fan of Glean, and that's that is a type of, I think I could, I could restate that what it is is a super -enhanced corporate, search product that helps you find things that have been scattered around the organization.

43:45Yeah, I think it's your enterprise assistant, right? I mean, Microsoft says they're capacity constrained because they got a lot of demand for co -pilot, right? And I think the real enterprise co -pilot, right? Glean has gone from enterprise search to being, you know, that enterprise assistant. I would say this, you and I have some debate about this, whether or not we're actually seeing the end user demand, whether or not the solutions are good enough to unlock that demand. I tend to be on the positive side of that question based on the companies we're seeing based upon the, you know, folks like Satya saying that we're capacity constrained.

44:21But I will also stipulate just as a risk factor in the public markets, the idea that we're going to go through this phase shift with CAPEX and revenues perfectly aligned and perfectly matched, right? It would be the first major phase shift we went through where that's the case. In the case of the internet, CAPEX got way out ahead of early revenues. In the case of cloud, CAPEX ahead of early revenues, mobile, same thing. So, in fact, I think what you're hearing out of these CEOs, once they're cautioning us, they're saying there's going to be a period of time here where we have really high CAPEX relative to our revenues.

44:59But it's the right bet to make when you look at a three to five year time horizon. And I would say on that dimension, each of those prior super cycles were underestimated when you looked at them three to five years out. And just to put a nail in the point I was making, I see a ton of great use cases and I'm super excited about everything I've seen. There are statements that are continually made that I think are over the top. You know, I don't think this replaces all software. I don't think people are going to be sitting around jobless in need UBI. I don't think you know, I don't think that all software just goes away.

45:44Someone said that this is the new operating system. I don't think there's any proof of that. Doesn't know how to store data. It may get there. But anyway, so the hyperbole causes me a little bit of cognitive dissonance. It's the only thing. But I do see a lot of positive stuff. Let's move on. Explain to people what happened with the Japan carry trade. Oh my God. I mean, you know, the fact that you wait. You show up on a Monday morning exactly. I thought I had to worry about taxes and recession and AI air pocket. You know, it wasn't on my bingo card. Maybe it should have been that we were going to have to, you know, be worried about a 13 % drawdown in the knee.

46:26Right. It's biggest one day drops since 1987. Then we have a rip back the next day. Well, let's break down what happened here. First, let's remember, the Japanese markets were really crowded, Bill. They hit an all time high just three weeks ago. High prices always make for big drops. Right. So there was a setup. But basically, the market was down big since the the Bank of Japan, so the Japanese Fed meeting where they basically increased interest rates for the first time in 17 years. So why was this such a big deal with a great threads on Twitter about this zero hedge has been basically live blogging the events.

47:07I think it's important before you go into what happens. Why were interest rates? Why had they not done an interest rate increase in 17 years? And why? Why did they feel they needed to? Right. I mean, so in Japan, you have, you know, a population that saves a lot, a population that's not growing a lot. And so they've been battling a slow economic growth and deflation for the better part of 15 years. And for the last eight years, they've had negative interest rates. Right. So that is just a way to force people into the risk pool to try to get people to take risk and get your economy going and to keep it from deflating.

47:51Remember, like in some ways, this is what, you know, we copied their experiment at the beginning of COVID when everybody went into lockdown and wasn't taking any risk, we said, we'll pay you to take risk and we took interest rates negative. Well, they're systematically been doing that for eight years. Now, it leads to one of these very interesting consequences, which is when you have negative rates, IE, you're paying somebody to borrow money. That leads to what is called a carry trade. So there are estimates that up to $20 trillion of yen was borrowed. And then you invest that in rate yielding assets around the world and you pocket the spread.

48:28Right. So remember, just I think earlier this year, maybe it was late last year, Buffett borrowed $200 billion of yen -denominated currency. So it seems like a great idea, especially if you're getting paid to borrow in Japan. So there's no cost to borrow in Japan. Then you get paid to buy bonds in the US. But you need to make sure that things don't change quickly on you. So the predictability of the rate environment in Japan is key. It's absolutely essential. And what scared the hell out of people is out of left field, you know, the bank of Japan raises rates. And all the sudden, these people who were effectively borrowing in Japanese yen, assuming that rates would stay low and investing in these higher yields, they had to unwind these trades.

49:20And you had 20 trillion of these trades at play according to JP Morgan. I think in the first two days, half of those trades they estimated got unwound. That means they had to sell US bonds. They had to sell stocks around the world, et cetera. And of course, that causes this negative reflexivity because you have the already half concerns about recession and the other things we've talked about. So this just adds to that list of problems. Now, remember, part of the reason that you have this carry trade going on is the US has kept rates higher for longer. So it increases the spread there at zero or negative.

49:59And our rates have been higher for longer. So that's led even more people to pig pile into this carry trade. But what was so shocking here, you know, within 36 hours, right? It seems like the bank of Japan capitulated. And overnight, they reverse course. And they said, we're just kidding, we're not going raise rates when the market is unstable. Now, what the hell does that mean? We're not going to raise rates when the market is unstable. You're the ones that destabilized it. So I guess that means sorry foot fault. We're not going to raise rates. But I think that, you know, most of the stock market contraction we saw here.

50:38So if you look at where the Nasdaq was trading before, you know, this all happened. I think it was like, you know, 475, it dropped to 450. And now we're basically back to 475. on the cues. So basically, it was one of these long tail events that reminded us there's risk in the world. And when you have a bunch of other buckets of risk, recession, slowing, AI air pocket, and then you throw one of these on it can really be the the the trigger to light the fuse. Remind you, just a couple of weeks ago, I can't even believe, you know, that that it happened at this point. But if it if that shot fired at Trump was one inch different, if Trump had been assassinated, can you imagine what what happened to the US markets?

51:26Right. So when markets are at all time highs and you're just hanging out and there's not a lot of upside return and you got these other things you're concerned about, you got to take down those units of risk because there are always things like this that can happen, right? Take around into account geopolitical risk. You know, we've come the world has come to I think appreciate the skirmishes between Iran and Israel over the course of the last decade. And they never really amount to all that much. And so when the markets get concerned about it, it's generally a buying opportunity because they are generally short lived, right?

52:03They keep each other in check. Could you imagine? Let's assume tail risk this time. What if Iran really did something different? And now we've heard rumors of you know, the Israeli leadership and nuclear proof bunkers. And they said, if you touch one of our civilian populations, we're going to wipe out your nuclear capabilities, right? Like you could imagine, I'm not saying it's the base case, but you can imagine a scenario where this time it is in fact different. I think for all these reasons, Bill, that's why you're seeing, you know, the markets taking, you know, just a bit more cautionary position.

52:41And to be perfectly honest, it's not off the table. This market could give up 100 % of the gains that we've had year to date, which would mean you still have another 10 to 15 % down in the market. And that just takes ship back to where you were on January 1. And so you have to ask yourself, as an investor, sitting here in August, is the world so much better than it was on January 1, right? That it has to be priced 10 to 15 % higher. And the answer to that is no. You know, my, I have a very similar thought on the, on the whole Japan carry trade thing, which is, you know, at the end of the day, it really shouldn't f***ing matter for American investors, right?

53:22This was an arbitrage thing that had been created by a systematic, you know, problem in Japan that was started with stagflation, but where they they had a problem. I think the reason they had to raise rates is they had a problem that the currency was causing. And they felt like they had to defend it. And, and, you know, the consequent was all these arbiters, you know, got wiped out. But there's no sympathy line for arbiters, right? And, and, and, and, and even if the banks got in trouble and went to the government, said, I need help because I had given loans to arbiters, they're not going to make whole, this is not going to be a long sympathy line for that either.

54:03And, but, but I do think what, what happened is, is exactly what you're saying, which is people were nervous already. And so like, you, you got your finger on the cell button. And then you wake up and you see this and said, shit, I said I sold last week. And you start pressing cell, right? Like it becomes self -re -imporcing, almost kind of like a reflexivity kind of thing. That's my, that's my non informed take on what happened there. No, I mean, listen, when, when people are off sides, you know, you know this hand in poker, right? Where you, you kind of limp in, maybe you have pocket sevens, right?

54:40And you got, you got a little too much money at play, but you're feeling good. You've been winning the whole night. And so you're, you're, you're, you're feeling a little swagger. That's kind of what a market is when it's at all time highs, right? You're playing the hand a little too strong. And then the flop comes and somebody has you trapped, right? And now you're like, okay, I got this sunk cost. Do I take it off the table? Do I, you know, do I bluff? Do I bet bigger? And, and, you know, I think this is one of these situations where people were just caught off sides, right? They viewed this as very, very easy money.

55:12And it's easy money until it's not. And, you know, and so, you know, I put all of these things in the category of healthy, right, resets in the markets, right? And so, you know, I do think there are a series of events like if we end up in a recession, if one of these situation, if Japan, you know, were, were to trigger something bigger, if the geopolitical events caused the price of oil to spike rate, all of those risks have to be taken into account, you know, particularly when stocks are at or near all time highs. And that just means, again, it doesn't mean you get out of the game, right? Like when prices are high in venture, you know, you don't just stop doing venture, right?

56:00But you do less. And you wait for the really great stuff. And when prices are high in the public markets, you don't necessarily step out of the markets altogether, but you just do less. And that's how we think about it here. It's lesser more. How many units have risked you want to have on at any one moment in time? And I agree. I do agree that of the of the topics you mentioned, the one that matters the most is the recession one going forward. So we'll see what happens. That's it. Two more things before we wrap up. The first thing we should dive into is this Google decision that happened this week, which is pretty landmark decision.

56:36I think there've been pressure on a lot of the Mag 7, you know, from the government and questions about monopoly status. And, you know, having seen Google from, you know, I met with Sergey and later when there are 25 employees today, I mean, kind of had a front row seat just to watch this amazing, you know, people have called it the best business model of all time. What do you take away from this decision? And what do you think it's going to mean for going forward? Well, I kind of want to turn the tables on you and ask you that, given how close you've been to it and how much you thought about this.

57:16But first, just like, let's talk about what the ruling in fact did. So, you know, it said that Google used unfair tactics, it's unfair tactics to maintain itself as the default search engine on the iPhone over the last decade. And it said it did this mainly by negotiating lucrative deals worth over $25 billion per year to cement its position as the default search engine on the iPhone, which then let it to have superior data, which then let it to have superior conversion, which then let it to further cement its position. I think the current deal is set to run through 2026. But the court, I believe, is set September as a date to kind of work out a solution with Apple and Google that the judge will deem to comply with the ruling.

58:06So, I guess the question I have for you is, you know, these guys negotiated a deal between the two of them. Google said, we'll pay you a bunch of money. Like, what do you think is the remedy? What do you think is a better alternative? You know, why can't they do a deal with Google? So, to your last question, you and I had afforded back and forth Ben Thompson's thoughts on this. And obviously, super intelligent human being Ben with a massive mental capacity. And he wrote a piece that will put it's subscription, but we'll put a link to it. Maybe we'll get him some new subs. But he broke down the Sherman Act, which was what was the piece of regulatory law that was applied here.

58:54And section one of the Sherman Act says, if you're a monopoly, and I don't think there's anyone that questions, you know, Google's monopoly status, just because market share, you know, north of 80, 90%. That's and and and Ben makes a point of saying they earned it through innovation versus some type of underhanded tactic or a roll up. And I think that's fair too. But but the Sherman Act doesn't say that there's a fair monopoly and unfair monopoly. It just says, are you a monopoly? And then section two says you can't enter into partnership deals basically that reinforce that monopoly position. And so the judge and Ben argues properly looked at the Apple deal for the iPhone and said, you know, this this violates section two, you're a monopoly and you're extending that monopoly with an exclusive deal.

59:53And so one thing I would back up and highlight for people, I think most people know this. But whatever this share is that Google's giving to Apple, let's say it's 30 % or 40%, who know, it might be 50. You know, it represents this massive amount of net income for Apple, right? What's the number? It's 26 billion, I think a year or two ago. So the reality is that Google's monetization is so much better than anyone else that's out there. That I don't think there's a deal you can do with being that will create that amount of net income for Apple. So there's no like Apple's not mad about this deal.

1:00:40They're very positively impacted by it. And so that's a that's a that's a that's a huge irony because they're not they're not. It's not a loss leader for Google where they're blocking someone out. It's a win win for both parties because Google's so good at monetization on search. That like I guess I'm making the argument if being gave them 100 % replacing them would cause Apple to have less money. So for sure, that's for sure the case. And what do you think the remedy is Bill? I mean, is this just going to end up in one of these it, you know, and I saw a screenshot of this, these sloppy choices where, you know, the next time I opened my iPhone, it says choose what, you know, where you what search engine you want to use for this query.

1:01:26And it gives me a choice between Bing and duck duck go and maybe now search GPT and maybe now perplexity and maybe now Google, you know, is that going to pop open every time? Do, you know, is it one and done? How do you think about that? And is that remedy enough like if the consumer chooses Google? And Google just pays a rev share rather than, you know, boxing everybody out. Do you think that remedy will be enough? It could be, it could be a nothing burger. So let's say that Apple is forced to do this election chart. Let's say Apple then tells every search engine in order to be a partner of us and in our search short, we'll take 30 % of whatever you earn.

1:02:10And then 90 % of people choose Google. I think you're in the same spot. I think you end up in the same spot. I think that's why I think it's silly. Yeah, go ahead. If they're forced to do the search and Apple somehow cut out of the economics and then Google ends up making more money, because I think the consumers are going to choose 80 or 90 % Google. So yeah, I think there are remedies where it could be nothing. One thing that would be super interesting, I think they're, I've often said that the federal government should have a new task for a monopoly. And this is a very self -centered bias to someone who's been a venture capitalist working with small companies for a long time.

1:02:57But for over the course of years, there have been many cases where I've been fortunate enough to work with a category leader in a vertical or something. And you get invited in by Google and Apple sometimes, sometimes it their request to do a deal. And you're presented with a set of terms that you would never consider with any other party out there. It's a completely one -sided deal. And I've often thought that should be a test for a monopoly. Like are they able to negotiate deals that are, there would just be off the table with any other set of partners? Because I do think it's a sign of strength.

1:03:41And I could even go further and say that it's a, that it should qualify under the section two of the Sherman Act. Because if the partnership deal results in the monopolist now having a new feature set because they were able to basically steal everything you've built through the terms of the deal, you've extended the monopoly through that practice. So that was one thing. I'm sure Jeremy Stopham would agree with but I just... That felt a little yelp like to me in terms of your concerns. I wasn't going to out you. No, not just yelp though. I mean, I think it's a fine line. Michael Porter's five forces would make the argument that this is what businesses should try to achieve, which is enough market power that you can extract rents.

1:04:36Commence it with your market power. And so I don't think we want to be casually undermining companies' ability to enter into commercial relationships. And those are going to reflect different market power in the economy. But to your point, listen, I think there's a pretty easy remedy here. I think the remedy is going to be consumer choice. I think the remedy is going to result in Google still getting 90 % of the searches. But, you know, it may not. And the only reason it may not is because we are at this really disruptive moment with things like perplexity and search GPT. And so listen, I'm happy to let them compete and to see where it all shakes out.

1:05:20And so speaking of, have you played with search GPT? I have not. Have you? No, all I've seen are the screenshots. You know, the video is everything looks very perplexity like to me. And just go ahead. I think it's worth stating. I think a lot of people know this, but I talked to a few people yesterday that didn't. I believe what happens at perplexity and why it's so much better when you're doing searches related to recent information. It's even good for the Olympics. I encourage people to use it is that in the background, when you type a search into perplexity, they're doing a search against a Google -like database.

1:06:07They're doing the equivalent of a Google search. Maybe I don't know where they built that. Maybe the partner would duck duck go or whatever. And they then take the results of that search and stuff it through rag into the into the context window and make it part of the prompt. And your result then is is focused on those things. And so it can handle newness, which everyone knows the original AI models couldn't do because they weren't trained on it. And that's also why a lot of people are like surprised perplexity has links. Well, the reason they have links is because of the process I just described.

1:06:47And so now, anyway, I think it's important for people to understand it now was I think the difference between search GPT and chat GPT is precisely that. They're now doing that same thing. Well, one of the things, you know, a few weeks ago, you cast some doubt on the engagement metrics, you know, at OpenAI. I mentioned I was a little more bullish on OpenAI, but I wanted to do the work. So I had a poor of on my team, you know, pull some metrics. And so I want to run through a few charts here about engagement and usage metrics that we're seeing on Gen AI apps compared to other consumer apps and enterprise apps.

1:07:27And, you know, it's get your reaction to this bill. So this first chart here is just weekly average users, monthly average users. And as you see, if you go to the far right on the chart, if you're like Spotify, Instagram, WhatsApp, you know, they have engagement metrics well over 80%. If you look at these generative AI apps, character tops out at about 64%, but if you look at clawed Gemini, chat GPT, they're about 40%. So much lower engagement, you know, then, you know, internet consumer apps. Now, if you go to the next slide. And by the way, just just just just so people understand, you're looking at, I mean, I guess is this MAU over?

1:08:17Yeah. Oh, it's WAU over MAU. So yeah, so weekly versus monthly. Okay. Correct. Correct. Now, if you go to the next slide, which this is, what does it take to get to a really big app? And if you go to the far right for those apps that have over 80 % engagement, right, they get to billions of users, but apps with much lower engagement, this all stands to reason. Those apps tend to cap out, right, at tens of millions, or maybe 100 million users. Right. So the question we were debating is can chat GPT get to a billion users? And like one of the early warning signs that you were talking about, this engagement doesn't look like it's the type of engagement that is commensurate with really growing your user base.

1:09:10So then we ask the question around retention. Yeah. Right. Which is, I think what you had specifically mentioned before. So here's month one retention. So this is one month retention. Again, if you go to WhatsApp, Instagram, Chrome, etc. Right. Retention metrics super high over 90 % retention. And again, enterprise apps, you know, like Gmail at 75%. But if you go to chat GPT, it's the best of the Gen AI apps at about 65%. But you see that these generate AI apps also have not only lower engagement, but they have lower retention. And you're mixing paid and unpaid on this. So it's tricky. Like you have to be careful.

1:09:56Duolingo's paid, tenders paid, Spotify's paid, WhatsApp X aren't paid. So anyway, that can be tricky. I think I think what we're trying to drill down on is, and this should all stand a reason. But low engagement, lower retention and lower usage as measured on this last chart by the minutes per week, you know, that these, that these apps get used. And what it tells me is this, Bill, and I just want to get your reaction to this. I love chat GPT. I love perplexity. You know, I love clawed. My team uses it all the time. Right. It's a better version of search. But the truth of the matter is Google's copied them pretty quickly.

1:10:44You know, it hasn't led to massive share shift. And it hasn't led to a step function in engagement, retention, and usage. And so I would say that search GPT, I don't think is going to be the product that gets up to a billion users. And it reminds me. It takes me all the way back to a conversation you and I had maybe an episode four, which is what is going to be the next, you know, GPT moment. And it feels to me like, you know, AI enabled search is not that moment. We maybe the moment is when we all have a personalized agent that has memory about us that can track us over time that really understands us that can take actions.

1:11:26We've talked about this on the pod. But I think this data that we pulled together kind of confirms the point you made a few weeks ago, which is barring right the launch of I don't know whether it's Q star or some other product, right. It feels like it's really this personalized agent that's a hundred X better in terms of driving my personal productivity.

1:11:54I don't even search GPT that are going to lead to the breakout needed to get to that billion users. By the way, this is this will sound a bit rude, but while you're talking, I did several search GPT searches. And it's exactly like perplexity. I think it's exactly what I described. I had it drilled down on the Christian Faulkner race. I don't know if you got to see that. It was fantastic. Yes, the bike, the this woman who was currently was briefly a venture capitalist, but the close she did was so brutal. Like she accelerated once she caught him. Anyway, I just had to walk through all that it lists the sources.

1:12:31It's very similar. I think there's some really big questions. The $20 thing is that real or not? Feels tough. Like it feels tough. It feels like Google's going to do it for free. It feels like perplexity is kind of cave and said they got to do it for free. We'll see. I think, but that's a big question because if you have to surround it with ads, it's a little different. And then the other key question, which we started talking about at the beginning of this podcast is if it is free, is the alternative to Google in having less ad units on it just better. And so do you do you is even if Google can lay chase, is it a situation where the disruptor just doesn't make as much money and therefore that puts pressure on them coming?

1:13:24I still think that's a possibility. I would pay the $20 not to have ads. There's a lot of questions about how big that universe of people is. I don't think it's the majority. And so anyway, yeah, it's interesting to watch. The one thing I would add to your one month retention numbers, the numbers I had done and I don't know if the service I'm using allows me to repost it. I'll have to ask them. But they had 35 % numbers for the 12 months retention, which is also pretty low and reinforces your point. The other thing I would last thing I would say about this is, I continue, when you start talking about that personal assistant, I continue to think apples in a decent position, but I really think that Google has, and it's interesting because I just talked about how Google may be disrupted negatively by this on search.

1:14:23But when I think about the personal assistant side, having the assets of Gmail, docs, spreadsheets, and Android, and I actually ran into Samir Samada, runs Android the other day and I had this discussion with them. It's the best set of assets. If you could pull it all together, it's the best set of assets. Because they have their hands on the mobile product, they have their hands on the email, they have their hands on your docs and infrastructure, they have a product that competes with slack in teams. So I could imagine if they perfect that integration with AI and what I like to call infinite memory or infinite recollection, people switching to their stack and to their phone because it's done so well.

1:15:16One big problem, Bill. One big problem, Bill. It's got execution. Google has been fumbling the ball and on getting there. They're going to be in the race. I would say when it comes to agents, the person who I think is talked most eloquently about it and is doing a ton of work is Zuckerberg at Metta. Open AI is going to be in that race. My main point here is not like we know all of these are going to be in the hunt. They have the capabilities to be in the hunt. Google may be advantaged from access to information and capabilities like you suggest. But what's clear to me is that the current state of the consumer product, it's not enough to dislodge Google.

1:15:59And that could improve. I mean, the memory part could could could dramatically improve improve. And and and I've said this many times, I think the voice things really cool. I finally got the upgraded voice assistant on chat GBT and I love playing with it. It's just so much fun. And and I'm sure it'll get better, but it does everything. I mean, you can just talk to it, you know, like you're sitting next to this incredible librarian. And it's fun to do it with other people around. People get a lot of kicks out of it. So we'll see what happens. I'm right. I continue to voice and and memory are the two things that I think are are the biggest and most exciting areas to add, but we'll see where it goes.

1:16:43Let's touch a one last topic before we we part ways. We had yet another one of these odd transactions that I'm going I wanted to brand it. So I'm going to throw out a phrase a take under rather than a take over. And so we've had inflection and and the one that just happened is character AI, which had pretty interesting engagement numbers on your on your slide here. What's going on? Why I mean, like, would it what's your take on these transactions? There's now been enough of them that it's I have to say it's a trend. Well, I mean, I think it's a combination of two things. Number one, I think from a company perspective, known Shazir as an incredible, you know, technologist engineer, visionary, you know, was one of the authors on attention is all you need.

1:17:41And he's building, you know, a race capital building a building a great business. You saw the engagement metrics. They're off the charts. But I imagine he's looking at the daunting amount of CapEx required, right, to continue to compete with meta and continue to compete with Google. And you know, that's that's a challenging Slav bill. And so on the one hand, you have these founders who are saying, I need to have I need to 10X my infrastructure stack in order to go, you know, attract the best engineers to work on the best projects, etc. And that's a full time job. And I'm not even sure you can get it done.

1:18:17On the other hand, Google and Microsoft and meta, they're willing to pay super big money for talent. But they can't acquire companies because we know we have a regulatory environment in this country. We're doing outright acquisitions as hard. And so I think they found this kind of third way that you describe. I don't exactly know how it works, bill to be perfectly honest. I haven't been in either of these deals. I would say that we looked at both inflection and character. We really liked what we saw at character. We wanted to be involved in it. But it was really hard to underwrite it for exactly this reason.

1:18:53We said the most likely outcome is that they're going to be a seller to Google or to or to meta because it's going to be very hard to break out. Now they're doing these deals. If it was just an outright sale of the business and you sold the business for 3X, what, you know, the series A investors put into the business. I don't even think we'd be talking about this. We would say, you know, it was a reasonable outcome. And they're going on and and and going to build something bigger and better, you know, at Google. I think what's interesting here is that these things are being structured not as straight up M &A in order to avoid perhaps some of the regulatory scrutiny.

1:19:32And so I don't know. I asked you do you like I'm just not sure how how they're structured or how they work. But the rumors at least as reported by, you know, the information and some other places is that, you know, investors are earning 2 or 3X on their money in character. I'm just not sure exactly how it structured. So I'm not 100 % sure either. And I've done quite a bit of research knocking on doors. And I also am not in one of those companies. And so I don't have perfect information. And if anyone does, if anyone wants to reach out and chat with me about it, explain to me how it works. I'd love to hear.

1:20:10But here I did talk to a few attorneys that have been around the Valley for a long time. And I would state this one. It is almost certainly the case that this the primary reason that these these funky transactions are happening are unless just called non -standard is precisely to avoid federal regulatory scrutiny. And so the minute that the Fed, the people in the Fed realize that's what's happening, they can start a process of laying chase, which might involve lawsuits might involve pushing on Congress to write rules in a different way. But they can't stop it right away. And so you're going to, you know, you'll probably continue to see this happen.

1:20:55It is, if it is an arb, the people that wrote the original rules won't be happy. Let's just state that like there's no reason they should. To, as I understand it, there is a substantial double taxation risk. And if this is not the case, it's a thing I'd love for someone to educate me on. But the way the money makes it from the mag seven to the investors, I think, is two step. One, the money is transferred to the startup as part of this massive license deal. And that's revenue for the company, which is exposed to corporate tax at that level. And then step two is either a stock buyback or a dividend distribution to the shareholders.

1:21:48And that two step process. And then obviously that's taxable at a personal level. And so that is, it seems like a sticky piece of this. And it means if someone truly is getting two X back on their money, they'd have to, the mag seven would be paying three X or whatever so that the taxes, you know, and I could, I mean, I'd love to hear how I'm wrong or I'd love to actually have perfect visibility into how one of these things plays out. I continue to doubt that it's preferable for the venture investor to a traditional deal. It just seems a little too sloppy and messy in this entity stays around.

1:22:39And people, it's funny in all these PR releases they try and posture as if this thing's going to achieve great stuff after all the core people left. And that that act of theater alone causes me to be a little skeptical of what's going on. Yeah, I mean, I don't know. I'd be surprised if this is the case with character. I think they had options. I think they had other buyers of the business. So I'd be surprised if it didn't work out in a way that was reasonably positive for the investors and reasonably positive for the team. You know, frankly, I think it was a bit of a coup for Google to get know him and his team back there.

1:23:15I mean, you saw these engagement metrics. They really are building, they have an eye to build something here that I think is important. So I like you. I look for a delirney more about this. But you know, and if they fix character and the things that I've read in the user forums and on Reddit is that it doesn't evolve with you. So it goes back to this memory issue and this kind of infinite recollection. But if they fix that, you know, this is a product that feels more at home at meta than it does at Google. So it'd be a new front for Google, I think, if they ever make this work. You know, I would say finally, Silicon Valley has been battling it out.

1:24:01We talked about it on last the last episode. You know, Washington Post talks about why Silicon Valley all in the Trump category. Now we have VCs for Kamala, you know, and in a long list there. You know, this seems to be the one thing that both sides in Silicon Valley agree on bill, which is, you know, Lina Khan and the regulatory environment that has been created that causes, you know, these companies to have to do these legal gymnastics in order to do acquisitions of small teams and small companies. It just seems that it doesn't make a lot of sense. Part of the beauty of capitalism is creative destruction.

1:24:42Companies come and go, companies get acquired, teams move around. That's part of the dynamism of it all. So whoever is in this new administration, you know, I certainly hope that we enter into a new period where, you know, they're enforcing the stuff that's on the books where monopolists are truly blocking out smaller competitors using their monopoly power. But we need to get back to a place where, you know, Microsoft and Google can go by, you know, a company that's diminimous relative to their size, you know, Amazon can buy vacuum cleaner company, you know, I robot without, you know, failings of lawyers and having to come up with these creative structures.

1:25:27I can't see how this is in any way good for the country. Yeah, understood. All right, buddy. Great seeing you. Thank you very much. Take care. We'll talk soon. Bye -bye.

1:25:46As a reminder to everybody, just our opinions, not investment advice.