M.G. Siegler: VC’s Crisis, Defensible Startups, and Hack-quisitions

29:08So it's not just fees. It's not just the money that was promised, but actually fees deployed. And so that is a little bit better of an alignment, right? Because then obviously you're only getting paid for the money that you actually put to work. There's a little bit of a weird alignment there, though, because you're then incentivized to put the money to work more. But anyway, I do think, though, in CRV's case, if that was the case and they were returning capital to their LPs, you know, it's probably a little bit less of, yeah, trying to, you know, make their own returns look better. Though, again, that comes into play.

29:45But it's also just like, yeah, telling telling the LPs like, look, we didn't deploy this capital. We're not going to take fees off of it. You guys just take it back. Yeah, that makes sense. Now, you know, one of the one of the things about this whole thing and you and you and you touch on this is that there's sort of danger and there's opportunity. Right. We just came out of this period. You mentioned Tiger. You mentioned the soft banks. And now like we're right back into this new period of AI where even the public companies are investing so much in AI. How do you think this ends up playing out for venture?

30:21Like, will there be exit paths to support the size of funds we're seeing? Or are we just like never we're just never going to get there? So, and that I think is the other key element of, of, you know, the article I wrote, because it does feel like even if you believe in AI and, you know, and I think I know you're certainly in that camp and I'm in that camp, it's not like the crypto thing. I do believe that there's like real, both interesting technology, but also interesting uses of it. And I think it's already sort of proving out usefulness in a number of different ways. But even if you believe that, there's a weird dynamic happening in that the M &A environment is just not really going at all right now.

31:06I mean, we talked about the Stripe acquisition, which was nice to see, in part because there has been nothing else that's happening. And that has largely been driven because of the regulatory fears from the US, but also in the EU. Any acquisition that's over a certain threshold is going to get looked at. Any acquisition that's in AI in particular, no matter the threshold, I think is going to get looked at. And that has caused a chilling effect on the entire industry. And so it means that whereas before the metas of the world and the Googles of the world and the Microsofts of the world, they would have naturally acquired, scooped up a bunch of these companies, some, you know, some I'm sure for, you know, multi-billions of dollars, but even just the smaller scale ones for them, at least in the hundreds of millions of dollars range, they're just not even doing them because it's such a headache.

31:53Like if they think it's not going to pass, you know, regulatory muster, it's not worth it for them to try to go down that path. It's such a headache. It ties up lawyers. It's just cost an insane amount of money to try to get a deal done. So we've seen sort of this. Yep. Breakup fees for sure. We obviously saw that with Figma deal. And so so instead, you're seeing these these what I'm calling acquisitions. But, you know, there's there's many funny, funny ways to frame them where it's basically like, you know, aqua hires that are being done and companies, quote unquote, licensing the technology, you know, of these companies.

32:31But really, they're just trying to obviously bring the team and the talent on board to be able to do these deals. And they're deals that in normal environments, I believe, would have been regular M &A deals. And those deals, I think it's important, like to underscore, those are very important to like the whole just flow and lifeblood of the industry, because it's everyone knows about the mega mega deals, the WhatsApp deals and, you know, multi-billion dollar deals, which are obviously, you know, great and much ballyhooed and are life changing for many, many people who work at those companies. But just as important, I think, are the smaller scale deals, which can also be life changing, obviously, for many people, but also just, again, lead to these sort of natural diasporas of then people who are within these companies who then, yeah, they made a decent amount of money, but not necessarily life-changing money, but they want to go out and found their next thing.

33:25And it also just creates like this natural exit point for a lot of talent to, yeah, go and figure out to do what's next rather than being, for lack of a better phrase, I mean, this sounds negative, but trapped inside of companies, right? Because they just, there's no natural exit path. Again, IPO window is shut. M &A window is not just is shut as well. And so what is the outcome there? The outcome is either a company going under, unfortunately, or people just leaving, which will probably just speed up a company going under. Right. And so it's a really, really bad situation right now. And the unfortunate reality is that even as exciting as AI might be, it's beholden to that because, again, unless regulations start to change and start to thaw and you have to believe that at some point, even, you know, slow moving bodies such as government recognize that that's an issue.

34:16But right now it's been like that for a few years. Yeah. Yeah. It's a good point about like why, you know, why these types of acquisitions keep, keep the sort of entrepreneurial spirit alive. I mean, you know, even the, the aqua hires, the small aqua hires, I mean, if you don't, if you don't have those, it can be, it can feel really risky to start a company, right. Or to join a startup, right? Because it's like IPO or bust, nothing in between. And when you do the math, it just becomes a lot harder to convince yourself to do that thing. And the weird side effect of that is it helps, it ends up helping the incumbent, the big companies, right?

34:52Because they know that both employees are less incentivized to leave, to go start whatever it is that, you know, that they want to start. Because again, it's just a much tougher environment out there, IPO or bust and IPO closed. So, you know, yeah. The flips, the counterpoints of this that, you know, I've heard a lot of people talk about, even people in tech who are incentivized to have the opposite, you know, say, well, at the same time, when the biggest companies in the world can do M &A, that also leads to incumbent advantages, right? We're talking about Facebook acquiring Instagram, acquiring WhatsApp, you know, many people, again, even within tech, talk about how Google or Alphabet should be broken up.

35:32I mean, how do you view like the flip side of all this? Yeah. And I mean, obviously, that's that's a lot of the stance that the government, the US government and the European governments have taken to date. I mean, my viewpoint, which is, you know, just guided by my own beliefs, not really any any, you know, super economic knowledge beyond sort of layman's terms would just be like what I've lived and breathed for the past 20 years is just like, I believe that markets naturally sort of sort that element out. Like, yeah, there are deals that happen, like meta acquiring, and I mean, what it was still called, obviously Facebook acquiring Instagram back in the day.

36:10Like, yeah, in hindsight, and we talked about the other one of DeepMind earlier, like in hindsight, obviously, you know, those companies should have stayed independent and they could have grown into the new behemoths, the new, you know, multi-hundred billion dollar companies. But the reality is too that like, That's obviously easy to say in hindsight. And you don't know, like, how much would Instagram really have grown without, you know, Facebook at the time's infrastructure and talent and everything else and helping to basically, you know, get it in front of, you know, hundreds of millions of more people.

36:43One that I was smaller ones that I was involved in back in the day was Vine and Periscope. So those were both acquired by Twitter back in the day. And it was very, you know, both were sort of a debate about like, you know, do you sell now versus keep going? Right. They thought they were tapping into something that was interesting. You know, obviously, Vine could not have foreseen what would come with TikTok, but it was basically TikTok in those early days. And, you know, there was all sorts of reasons why Twitter might have screwed that up. But regardless, you know, it's like the decision ultimately came down to do we believe that we can go farther with the acquiring company?

37:24And, you know, you could even speak to that better than than anyone. Right. As as as someone who had your company decide to go down that path with Spotify and, you know, would love your take on that. But I think, you know, when we were when we were discussing various options like that was a debate, right? Yeah, no, for sure. I mean, look, I have no regrets about Anchor being acquired by Spotify. I think that the podcast ecosystem benefit from it. After having spent years in the podcast ecosystem, I don't think Anchor could have survived independently long term for a number of factors that would probably take up a whole podcast.

38:03So obviously I feel good about it. At the same time, when I see deals that come to me in the podcasting space, I look at them, you know, with an initial wave of sort of apprehension because I'm like, oh, well, that market, the market's already sort of cooked, you know, like we sort of know what that's going to be. Very hard for me to imagine that a company will come and disrupt, you know, Spotify and YouTube in terms of podcasting. So it cuts both ways. It's really, really complicated. it's not black and white it does it does i agree with that but i again i my mentality just goes back to i think that the government really is messing up with oh i totally stopping the this m &a uh environment now certain types of m &a sure like you know massive method like should google be allowed to acquire another search company like you know these types right okay that's one thing but like even like amazon and the the robot vacuum one robot that was bizarre that was so yeah like And Meta acquiring the VR fitness one.

39:05Like, why are they spending time looking at this? Yeah, what are we doing? Exactly. Oh, man. Yeah, it's crazy. So what do you think happens as a result moving forward? Do you expect we're going to see more funds returning part of their funds? Do you think on this next sort of fundraising cycle from VC funds that we'll see smaller funds? Or do you think we're just going to keep, you know, the trajectory of bigger and bigger funds is just going to continue? It all, in my mind, at least, it all ties into exactly what we're talking about. If the IPO window opens again, if M &A starts happening again, and sort of in particular within the AI world, like things start to thaw, exits start to thaw.

39:51And then I think that it keeps sort of the cycle supercharged and going. I think there's a number of other things at play, though. Like if that doesn't happen, I think it becomes a lot more complicated quickly in the next, say, two to three years from a fund perspective on down. Like, do you keep investing in companies even when there's no obvious exit potential? The answer would probably be no, but there's still going to be opportunities. I would have to imagine playing this out of my head because there will be like the people who start to leave open AI. I mean, there's already, as everyone knows, a lot of people leaving open AI.

40:30But as more and more as the company gets bigger and bigger and, you know, and continues on its path, there will be more people that spin out. And VCs will obviously be fast to fund a lot of those people, sort of no questions asked type stuff. Same within Google, within Meta, within all of the companies, there will be AI talent that is still spinning out. And so VCs will jump on these would be my guess. but from yeah just the pure pure new startup ai plays i think that they start to slow more again naturally and i think that that's probably a healthy thing but the funds themselves again it all depends on what the exits sort of start to look like and if there's opportunities to get money just flowing back into the system because remember we talked about like yeah like It's good for diasporas and people leaving companies and figuring out what to do next.

41:26A lot of money is just locked up on the investor side in various companies that they haven't been able to return capital to LPs. And it's leading to all sorts of different other dynamics within the market of like secondary sales and different SPVs and all these different things that are coming into play in order to just free up some liquidity on the investor side as well. And so just a lot of unnatural dynamics that feel like they have to play out. And again, if I had to guess, if I had to extrapolate out how this goes, and I would love your take on this from the opposite angle. But I think that from the top down, the biggest companies, I think you're already seeing it, are starting to sort of slow in terms of where the next models are coming, right?

42:12Like it's now sort of being reported that like GPT-5 and whatever Google's next Gemini is and all these things are sort of like being a little bit elongated out and coming into next year. And it's taking such immense compute resources now to actually make these things come into play. Like, is there some sort of new breakthrough in chips or whatnot that actually re-accelerates it? Or are we sort of going towards the tail, you know, the long, slow sort of move into like, this is where we are. And it's just, you know, we're going to keep making breakthroughs in AI, but they're not going to be what they have been for the past two years.

42:51And so if that does slow to sort of those points, I think that that has trickle down effects on the entire ecosystem. And I think in some ways it'll be very good because it will allow things to slow enough where real companies can be built on top of that technology. But at the same time, I think it changes the funding dynamic of what those look like. I think in terms of sort of pure scaling laws with sort of the existing architectural frameworks and the existing data sets. Yeah, like we may have we've sort of like it seems like we've tapped we've tapped all the sort of natural data that exists, at least in the digital world.

43:30We haven't explored sort of the physical world. but I know that there are a lot of really smart teams working on algorithmic breakthroughs which I think many people believe would just keep this thing it would just send this thing further further up and I know there are a lot of people obviously really smart people thinking about obviously synthetic data and new sources of data from kind of like the physical world and so you could see it going both ways I think is the answer and yeah I think you're going to continue to see a lot of capital follow some of those follow those some of those trends for sure um but it's hard to know and and like you said if there are no exits and there's no path to exits what do you do i mean one of the interesting things is like yeah maybe there are fewer uh companies that can seek exit paths at the same time the cost of fund this stuff is still so expensive so those two things like almost cancel each other out in some way right where it's like maybe it's just more concentrated in terms of the bets.

44:33But then you obviously need really, really, really big outcomes to pay off those bets. So when you're investing now, what do you look for in these? When you meet a new company that's sort of in the AI space, like what is it that you're looking for? Yeah. I mean, I think number one, something you said to me a long time ago that has only recently crystallized for me is first of all, people like not that, you know, didn't value people really early on. But I think like when I was a very new investor, I tend, I tended to gravitate more towards ideas. Obviously I still place value on ideas, but now I place even more value on people.

45:09So obviously people look, I think given the new capabilities that are being afforded to us as a result of AI, just like always, I'm looking for great products that have the potential to accumulate users, retention, revenue. of course, in a way that is hard for incumbents to follow. And I would say with that last piece, it has probably less to do with AI because obviously, as you mentioned, all these incumbents are investing in AI as well. So it can't really be about AI. It actually has to be more about the same things that always made startups defensible and have a moat, right? Which is like either a unique data advantage or network effects or a first mover advantage.

45:57that locks in a user base very, very, you know, in a very, very sticky way. The other thing I'm thinking a lot about is sort of the new business models that are being created as a result of AI. So a recent investment that we announced was one in this company called Tolbit. I don't know if you've seen this company, but, you know, sort of all these AI agents are sort of increasingly and we think much more increasingly will lead into ad revenue because we're getting less sort of human eyeballs visiting web pages. You know, Tolbit is trying to figure out, okay, how do you create a new business model and compensation structure for publishers in a world where the existing business model of the internet starts to fade?

46:40So I think there'll be stuff like that. And then, yeah, truthfully, also the foundation model layer, which we just talked about, I think that tends to focus, again, on companies that are seeking new algorithmic breakthroughs, not just taking the sort of traditional scaling laws of transformers and just throwing more data at them, more compute. And then maybe some companies that are taking some new and unique approaches to data that they could feed more of the traditional architecture. So it's a lot of things always. Are you a believer that we get to AGI? Are you a non-believer in that? I think it's so hard to know what AGI even means.

47:24But I do think that we, I do believe that we will get to a point, even with the current scaling laws and the current architectures and the current sources of data, that we will get to a point where these things are much more intelligent than humans and potentially, you know, infinitely replicable. at least within, you know, the parameters of like cost and compute and that sort of thing. And I think, you know, if you assume these things are agentic, that creates like a pretty interesting world, a pretty interesting future. I'm not necessarily saying that like robots are going to come and kill us tomorrow and, you know, find a way to make us extinct.

48:06But I do think there's going to be probably some pretty massive disruption to the job markets, especially for, you know, things, tasks that probably could be handled by, you know, sort of an average digital knowledge worker. Again, I don't know if that's considered AGI or what, but to me that feels like that's coming. What about you? Yeah, I mean, I think I'm roughly in the same camp. I think the biggest problem I have is with the definition of it because it's the moving target constantly. And so it's like, what does that actually mean? I do think to something that you brought up earlier, you know, and again, sort of in my layman's understanding of this, I do think like we will hit that wall with LLMs and you need to figure out either if it's a new algorithmic breakthrough or like real world data, I think, yeah, it's become super interesting, right?

48:59Is it's not just like, yes, we've tapped into the entire written, you know, works of humanity and that's amazing. But there's still, you know there's like is there um is there like varying degrees of nuance and yeah like how you look at something and you know all these all these different things which are part of the human experience which how they would play into uh actual intelligence yeah we learn from more than just reading right right of course we learn from looking at things from hearing things from hearing intonation of people's voices you know i think as these things become more and more multimodal and they get distributed to more and more different endpoints, there's naturally more real world data it can take in and learn from.

49:43So what's your view there? Because that's like a natural, you know, segue to talk about the basically new devices, like our new devices needed for AI stuff. Because, yeah, I mean, my general viewpoint was like when the humane and rabbits of the world were coming out, right? It's like there's interesting things about them for sure. I think their timing is totally wrong. And I think that they're too complex to really make it in the way that they sort of came out to market. And some of it was PR and bad branding or whatnot, but it also just felt too early. But I do think that those types of things are interesting.

50:18And now, of course, what we're seeing with Apple, while I think their early stuff is a little bit too rudimentary, as I think the point has been made very much so. But I do think the visual intelligence stuff that they that they can do because they have a billion devices in everyone's hand like and they can tap into that whereas like who else can do that i mean obviously google i guess with with android but even they're like are they going to be able to fully tap android or can they only do that on pixel phones and then well you know what does open ai do with whatever the johnny ive thing is and like um like what's what's your viewpoint on how hardware comes into play in those yeah i mean i think there's there's probably like way more uh mileage these companies can get out of the current form factors to feed into you know ai sources of data for ai right like i have to imagine that the capabilities of the watch get more advanced as these things get llms on them same thing with the phones obviously same things with the airpods you know i know there's and talk, oh, will there be cameras on the AirPods?

51:25Maybe. Obviously, Orion and the glasses that Snap have been working on, you know, less so Apple because we haven't seen what like a more organic and useful device looks like from them. But I think like I say that because I'm not convinced that a new form factor emerges unless it's so obvious and convenient to users for why they should or people for why they should use it. Right. Like you mentioned some of these new devices. I don't think people just pick up and adopt new devices because they're new and cool, right? They adopt them because they're really, really useful. And they solve something for them in terms of convenience or entertainment.

52:03And that was the smart framing of both Snap and Meta in their cycles, talking about their AR. They both make the point. Evan Spiegel and Mark Zuckerberg both made the exact same point, which is like, people already wear these, even if you don't wear glasses, you wear sunglasses at times. And so it's like you're not trying to make them wear pendants or make them wear something that's like, yeah, that they would not do. Yeah, I think we're going to see these companies lean into the form factors that we already have and have adopted to find new and new ways to wedge in AI, to collect more data, to then obviously get smarter.

52:37But then I also think like it'll come from other devices and other interfaces as well. Again, maybe not in terms of four factors that we as sort of human beings walking around adopt, but, you know, from our cars, as an example, obviously, like my Tesla has however many cameras on it. I'm assuming more and more of that data will be sort of ingested and adopted. You know, you think about a few years ago with this like Internet of Things, how there's like, I don't know, AI in my refrigerator or obviously like the smart speakers. Like I think like these companies are going to lean more and more into the devices that we've already adopted.

53:19Maybe maybe refrigerator is a bad example. But like, I'm sure the speakers in the home, there will probably be a reinvestment in all the speakers or devices, smart home devices. There'll probably be a reinvestment of all those things. Even the refrigerator example, as silly as it sounds like, again, going back to like, that's all interesting data for something, right? Even if it's not explicitly used, like it's going to be, it takes into account different things. And you never know if there's intangibles that that helps out with in some training of some algorithm. to help all this do better. The one last thing that ties into that is, what's your current view on consumer startups?

54:00I get asked this all the time. And so I'm assuming you do. I mean, I know like it's trendy to say consumer's dead, but I completely disagree. I've never agreed with that. I think like certain opportunities are gonna be really, really hard to break into. Maybe ones that were much easier to break into eight, 10 years ago. So I think oftentimes when people talk about consumers dead, it almost seems like they're referring to social media, right? Which I'm not even willing to say that. I think it's obviously way, way, way, way harder to build and scale a social network effectively in 2024 than it was maybe in 2014.

54:41Obviously, the current players have these graphs that are impenetrable and these algorithms that are really, really good. You can't get them unless you have a bunch of data. So like maybe traditional social network is going to be really hard to do. Counterpoint, like we've got AI now. AI can do things that we never really could do before. So maybe there's going to be some interesting and unique form factors that emerge that you and I haven't thought of yet. But I'd say like consumer more broadly outside of social as a category, super bullish on. My take is if this AI really gets agentic and is really sort of augmenting all these workflows for us, especially in the workplace, Like we're probably we probably will have some incremental time to reinvest.

55:24And like, why wouldn't we invest it in the things that bring us entertainment, joy, health, happiness? Like. We spend our time on the things that we like to do as much as possible, we would like to not do the things that we don't enjoy. Right. And I think to me, that's just much more synonymous with consumer than it is enterprise. And so I'm bullish on consumer, especially in sort of like a post-AI world, what that means exactly. I, you know, I'm sort of, when I say that, I'm being sort of category agnostic. I don't know. What do you think? I think that that's, that your point is a really good one that I haven't heard sort of phrased that way, which is that it's something I thought about a lot in regard to, you know, going back to the last cycle again, where it's like COVID times.

56:15Looking at sort of the consumer landscape, you could see why exactly the rise of Clubhouse happened, right? Because all of a sudden you're locked, everyone's locked at home. You have this time that sort of opened that you have nothing but time. But it opened and it opened in sort of this unique way, which was uniquely advantageous to Clubhouse because we had these devices. We had AirPods and right. And so everyone's sort of used to that. And the time unlock was the key, though. And then COVID went away and the time went away. And so it's like, that's unfortunately how that played out. But with these AI changes that you're talking about, if it does unlock time, you know, time saved from doing rudimentary work or whatnot, that's potentially going to be unlocked and doesn't go back.

57:02um you know it doesn't go away you don't lose that time and so there's there's pockets of time that are freed up where there's things that can fill that that void now all of a sudden and what does that look like and i think that that's super interesting for sure i also think um you know entertainment media as categories i think that stuff gets i think that stuff has major tailwinds one for the reason we're talking about like we as humans we like to spend our time uh entertaining ourselves, watching Netflix, scrolling TikTok, whatever. But not only that, we just talked about a few minutes ago, you know, I mentioned this idea around the business model of the internet changing as a result of agents.

57:43You know, we also, there's also this dynamic where there's probably going to be an abundance of content as a result of generative AI. And so the best stuff, I think there'll be a premium on it, right? Like the best content is going to be even more valuable because you want to invest even more of your time in it. And the best stuff will be sort of rare, right? It'll be rare compared to all the slop and junk that's out there. And so, yeah, look, I'm bullish. I'm consumer. I don't know where it goes. I'm, you know, but I think I think I think saying anything in like a really binary way right now is probably a mistake, like consumers dead or social media is over.

58:28That's how the internet works, Mike. You gotta. Yeah, exactly. No, but I do think that's a very good point on consumption and creation of new types of content, too. I did write another post maybe even longer ago than that. But it's like, I look at the fears around Hollywood and generative AI and stuff. And I just like that's in my mind, at least that's going to play out the same way that all technology coming into Hollywood is played where it's like, yes, it's scary at first. And then and it and it will display some level of jobs or change them at the very least. But it will end up being such a benefit to that entire industry.

59:10And I think the most important point is the one that you made. It will also raise the value of human creation higher because it will be valued, you know, as a more bespoke experience, you know, versus the full on generative stuff that comes out. And so it will actually help people in those industries in ways that they just can't possibly see right now. 100 % yeah I totally agree I feel I would be remiss if we didn't spend at least a little bit of time talking about Apple you are one of the world's uh I don't know leading opinions of Apple I feel like that's just I don't know about that but if there's anything it's just from being a reporter back in the day covering it for years and years and so I've been on like the Apple beat for 20 years or whatever of course yeah so we got to talk Apple and we got to talk apple intelligence uh if we're talking ai maybe just to start like i would love your high level take on where apple stands as a company with respect to ai are they behind are they ahead um are they right on time and then we can get into some of the more specifics around how is apple intelligence what is this going to look like etc etc so my general view has been before they launched any of the new stuff recently.

1:00:34I think that from a technical standpoint, my belief, and again, I'm not inside the company, so I don't know, but my belief is that they're behind. But I think from a product perspective, they're not necessarily behind. I think that it's still so early with things moving so fast that I think their strategy, while I think it could use some tweaks in terms of like the stuff that they initially rolled out, they probably could have rolled out more, you know, and it sounds like they're going to fast follow with some other stuff that looks a little bit better soon in the 18.2 beta. So I don't necessarily agree with their exact cadence of it, but I think at a high level, conceptually, what they're doing is pretty good.

1:01:17I think about, yeah, parents. And, you know, again, it's a billion plus people using these devices. Like they don't necessarily want, yeah, like some sort of, you know, wild auto-tune, you know, something in their voicemail inbox to be able to have fun with and make AI generative stuff in there. But I think that they do like the idea of, you know, like notification summaries and things like that that are actually tangible and usable. Some of the work there needs some work, but you get what they're doing. And so again, I think that they're not off on their timing in terms of bringing the products to market.

1:01:57But I do worry that they are behind in terms of technical talents and capabilities if this stuff starts to move faster. So just as we were talking about earlier, if there is indeed this sort of slowdown happening in LLMs and we get to the point where you can build all sorts of stuff on top of those much more quickly because there's not the ground shifting under your feet all the time in terms of like capabilities, I think that, you know, Google and Microsoft OpenAI are situated very well to quickly pounce on those and roll out really interesting stuff. Whereas Apple, I think, will be too cautious, would be my read of how they would do it.

1:02:37But again, that that is a certain world. So I think that they're situated fine for now. But again, if we get to a point where people are sort of looking for more advanced uses of AI quickly, then it becomes a little bit more problematic for them. Yeah. Did you find it strange that they shipped the new iPhone without any Apple intelligence in it? To me, that is an indication that they were pressured by, and they would never admit this, but that they were pressured by Wall Street into having answers. They look at their peer group. They look at what happened with Microsoft stock, thanks in large part to the OpenAI investment.

1:03:20They look at what happens to Google stock. They look at what happened to Meta stock, all because it's all AI all the time. That's what they're talking about on these earnings reports. And so if they didn't have a credible answer for that, and look, you could take the cynical viewpoint of that entire situation, but it's also just like reality. Apple is one of the most important stocks in the world because it's owned by mutual funds And like it's in everyone's 401k, right? It's like, you know, it's a huge portion of Berkshire Hathaway. They have a reality of the size that they're at that they have to have credible answers for that type of stuff.

1:03:53And so it's in their fiduciary responsibility in some ways to do have credible answers for what their AI answer is going to be. And so, again, I think that they rolled this stuff out before they were ready. I think in a perfect world, they probably would have waited another year. We probably would have heard about it at next year's WWDC, and it would have been much more fully baked. And the devices obviously would have had it from day one. Because, yeah, they just look sort of silly by doing all this marketing for all these features that aren't even out yet. And a lot of the Siri stuff sounds like not even coming until well into next year.

1:04:26Yeah. What do you make of the partnership with OpenAI? Obviously, that was part of their keynote, I guess, at WWDC in June. they were supposed to invest they didn't yeah so what do you make of that what's going on there i do think again i think in part you could probably chalk that up to the you know notion at least from wall street's perspective that they were behind and how do you get sort of up to speed really fast beyond what you're doing internally you can partner and apple has a history of doing this right like they they bring they brought on um google to do maps within the original iphone right and and youtube and and they have and yahoo for uh for stocks and all sorts of stuff so they've done this historically and then of course you would imagine that as time goes on as they build up the muscle internally they go and roll out their own version of you know what they were partnering on previously um but this with open ai in particular is interesting because open ai is a very interesting company as everyone's very well aware at this point um it does feel like maybe Maybe they, you know, thought, you know, wanted to do a bit of a bake off and see, you know, what which they should use.

1:05:40And, you know, there was talk of potentially using Google and maybe that still comes down the road with Gemini stuff. But but it felt like, you know, open AI was probably the right one optically for, you know, from a Wall Street perspective for them to use. And then I was always curious in those reports. And again, no real knowledge, inside knowledge about any of this stuff. But it felt like it would be very strange to me if Apple was doing that type of partnership and if OpenAI and Microsoft was OK with that type of partnership. If Apple was just totally freeloading on OpenAI, because when they turn that on to a, you know, it won't be a billion devices because it's going to be relatively muted in the U.S.

1:06:18and blah, blah, blah. but when they turn it on to tens of millions hundreds of millions of devices that's gonna that's gonna fry open ai servers even more so than it already has um because i don't know if you've used it at all in the betas like it's in the newest uh ios 18.2 beta oh uh the chat gpt integration is live it it defaults to it pretty quickly it's not like it's not like oh uh yeah you have to like you know summon it via some special incantation there is ways to do that I think you could say like, hey, chat GPT, and Siri understands that you want to talk to chat GPT. But it also like if Siri doesn't understand something, she will quickly flip the switch over to chat GPT.

1:06:58And, you know, they obviously, as you know, prompt you to say like, would you like to use chat GPT for this query? And so assuming that that happens at the same sort of scale as my limited usage so far, again, it's going to hit their servers hard. and those are Microsoft servers that Microsoft has invested in OpenAI to pay Microsoft for and yada, all the circuitous stuff going on there. And so I would have just been very surprised if there was no financial agreement and obviously it was reported that they didn't have financial terms in the early days. But again, that's why when it was reported later that Apple would be investing, that to me made a ton of sense because it's essentially a way for Apple to pay for the chat GBT usage.

1:07:39But they're not now, which is obviously a surprise to them and everyone else. At the last second, maybe because of the Miramarati stuff, maybe some other reasons they got cold feet. But so that whole dynamic is just really strange to me now. And I have no idea how that ends up playing out. It seems like, you know, obviously they're going forward. Again, it's in the beta with the ChatGPT integration. Do they bring in Google or some of the other or Anthropic or someone else faster to try to offset some of that? I don't know the answer to that. But that will be really fascinating how that plays out over the next like six months when they roll out ChatGPT.

1:08:12Could you see them bringing in other models? And what would that experience be like? Would it be a setting in iOS where you sort of pick your default? Yeah. So it is already like because you can go into the settings in 18.2 beta right now. So they have this new thing that they rolled out. It was ostensibly, it sounds like built for the EU because of the new mandates that you have to set new defaults, right? Like you can set a new default browser and a new default messages app. But they also have in the settings a way for you to log into ChatGPT if you have a premium account. And so you could easily see a world in which they just basically have, you know, you set your preferred LLM for that purpose.

1:08:58And yeah, so Gemini or Claude and, you know, all the other things. And maybe they even like just have those as payloads in in their individual apps where you could do it that way. And is there is there any sort of in the beta? Is there any sort of like cutoff that pushes you to a sort of like chat GBT paywall? That's a good question. As far as I know, no. And that's, again, why I think, yeah, this is now maybe they implement something like that because it feels like, yeah, you could easily that could easily spiral out of control. Like, yeah. um so maybe they yeah like but as of right now no there's no like you don't get like 10 free queries to just let you go yeah it's fascinating i mean back to our earlier conversation it feels like if there was ever an opportunity to make an acquisition in this space which maybe they can't for all the reasons we discussed apple would be wise i think that they will acquire something here um maybe not open ai maybe perplexity maybe who knows my bet would be that they will um they will if they can how can they though right well you're right so and i think like the information is is written they wrote specifically about i think they just threw out an idea i don't think it was like a reporting element to it but it's like should they buy mistrawl i think they were like yeah throwing out potential m &a in next year for ai it's like mistrawl is interesting because it's in europe um you know it's it's open source in the same way that i guess llama is so you know open weight more um but uh it's a smaller team um and would they be able to you know spend overseas cash on it and and bring in a team there um again maybe that's easier maybe that's harder because it's in the eu and they have their own you know war going on with with that with those entities but um i would i would bet that they i think that one of the next big acquisitions that they make and they only make very rarely right they made like beats and they made siri back in the day and and um uh you know a couple others like the one that ended up with that became apple silicon um i think that they make a an acquisition in the ai space basically for the reasons we talked about earlier to to fully catch up and yeah something yeah well we shall think amazon will too they seem to be closely tied with anthropic anthropic yeah that's what I've heard.

1:11:22Yeah. Trying to get that through regulatory. Good luck. But yeah, you can't get a robot. You can't get a robot vacuum acquisition through. MG, always a pleasure catching up with you. We got to do it more often. Thank you so much for coming on the podcast. And I learned a ton. I'm sure the audience did as well. So really appreciate it. My pleasure. It was great to chat and we'll have to do it again and come to London. Come visit. Let's do it. Let's do it. Awesome. Thanks a lot. Thank you so much for listening to Generative Now. If you liked what you heard, please rate and review the podcast. That really does help.

1:11:58And of course, subscribe to the podcast so you get notified every time we publish a new episode. If you want to learn more, follow Lightspeed at LightspeedVP on YouTube, X, or LinkedIn. You can follow me at Magnano, M-I-G-N-A-N-O on all the same places. And Generative Now is produced by Lightspeed in partnership with Pod People. I am Michael Magnano and we will be back next week. See you then.