Why Lightspeed Bet on xAI, Neuralink, Suno, Granola, & Pika | Michael Mignano

[00:00] This is like a very hot topic. I'm sure you have a lot of opinions. What did you think about the Sora launch? I think Sora is the next phase. It's a whole new chapter for the internet. You know, I hate to say it. It's almost like the end of the creator. Okay, so you joined Lightspeed, $30 billion fund. You've invested in XAI, Neuralink, Suno, Granola, Pika. So I'm curious, how do you see creativity as a moat? When you think about moats, I think what's really important right now, [00:30] sticky and durable extremely early. So one lesson for me is Suno. It's producing the best AI generated music in the world. They have a shockingly high amount of revenue that I will let them reveal on your podcast. I won't do it. But the thesis has been extremely validated. And I think you only get there by being creative and having the courage to go do it. What do you think the world's going to look like once we have super intelligence? Uh, well, I don't know. [00:58] *outro music* [01:06] All right. All right. Michael, welcome to Sorcery. Thank you, Molly. Happy to be here. As the godfather of podcasting, how does it feel? It feels good to be recording a podcast. I like podcasts. Did you expect it to be this crazy? No, absolutely not. When we started Anchor, it was a very niche thing, I would say, in the Apple Podcast Store, which Spotify wasn't even really a podcast platform at that point. It was just music.

[01:36] in the store. [01:38] And then, you know, we didn't even think it was going to be big. So we were actually trying to build a social network for audio. Like we imagined like really short form stuff, kind of like Clubhouse. [01:48] But then eventually, after a bunch of pivots, we just realized that, oh, wait, podcasts actually are a thing. We actually have to attach ourselves to this. We need to make this really, really easy. And then the whole thing just took off and ran away. And, you know, then there are tens of millions or maybe hundreds of millions of people that have now made podcasts. So... [02:06] It's crazy. We did not expect it. And you were head of talk. That was the title? At Spotify, yeah. I think that was my title, because talk audio encompassed podcasts. We were also doing video. This was the start of Spotify getting into video podcasts. And then back then, this was a few years ago, of course, we were also dabbling in live talk audio as well, because after Clubhouse, everyone had to have a live play. The problem with all these live... [02:33] formats on the internet is that synchronous content is really, really hard because the beauty of the internet is that everything is actually asynchronous. So you can consume it [02:44] on your own schedule, right? Like this podcast or all the podcasts you've done, anyone can consume them on demand. But if you have a synchronous platform, you're sort of forcing everyone to say, "You must consume this at this exact moment." [02:56] And that's just really inconvenient. [02:57] So fast forward in 2022, you joined Lightspeed? Yes. Okay, so you joined Lightspeed, $30 billion fund, a mega fund. We had Bucky on. It was great. Oh, nice. Bucky's the best. He's the best. Yeah, we went into the economics of what it takes to return a $30 billion fund, which was really interesting. And,

[03:17] I think we could relate this to some of your portfolio companies. So for you, you've invested in XAI, Neuralink, Suno, Granola, Macroscope, [03:26] Pika? [03:26] So you've done all of those. And then I've also read on you that creativity is your moat that you like to go after. So what is your investment thesis? My investment thesis, when I started, candidly, I thought it was going to be really focused on things that matched my own experiences and my background. You noted that I built Anchor. I spent a lot of time building consumer products for lots and lots of people. And I love those types of companies. [03:56] products for everyone, [03:58] But the reality is, after I joined, AI became a thing. And I think with each day that passes, I think we're realizing that [04:07] AI makes it so that we know less and less about the future. I think every investor out there has some thing they're good at, or some way they see the world that's different than other people. But I also think in many ways, [04:20] we're all having to forget kind of everything we know and dream a little bit bigger because things are moving so, so fast. But for me, yes, I am drawn towards media. I'm drawn towards creativity. And at the end of the day, as like a product builder and a product designer myself, I'm drawn towards founders who just have amazing product taste and sensibility. You know, you mentioned Granola, the founder and CEO of that company, Chris Pedregal, with his co-founder Sam. [04:47] I had known Chris for a really long time, and I watched him build this company, Socratic,

[04:52] And I just knew that this guy had incredible product intuition. Just like, I knew whatever he chose to build, something really great would come out of it, because he had shown me that through his previous company. And so when he told me he was building something new with Sam, it was like, [05:07] doesn't even matter what you're building, sure, we will invest. [05:11] And so I find more often than not, especially as AI, again, kind of makes you have to forget everything you know, it's really becoming more and more about people, great people. And for me, it's like product oriented CEOs. [05:24] How does XAI and Neuralink, those size investments, fit into this? Yeah, it's a good question. So maybe taking a step back, Lightspeed, as you mentioned, it's a large platform. Mega. 30-ish billion under management. The current fund generation we're investing out of is in the ballpark of about 7 billion. [05:44] To do that, we're basically investing across stages. And I generally focus on early stage, but we're a firm that really prides itself on collaboration, working with our partners across different stages and sectors. [05:59] If somebody has a great relationship or if somebody has a very specific point of view on the world, we're not a firm person. [06:05] that lets the borders of those sectors or stages be a barrier that prevents a deal from getting done. And so I generally focus earlier, but as a fund, we absolutely have a strategy to bet on the biggest and best companies in the world. And in the case of companies like XAI and Neuralink, it's like you got to back Elon Musk if you have the opportunity to. And so,

[06:27] That's the idea behind those investments. But for me, I am more and more drawn towards the earliest stages. Again, as AI is making it such that you kind of have to [06:36] forget what you believe and [06:39] when you're living in that kind of world and operating in that, you know, in a world that moves so fast, the best bet you can make is on people. [06:46] And so right now I'm doing a lot of seed, a lot of like inception stage. This person's amazing. Let's give them a couple million bucks type of investing. I'm really interested to know how you think about this environment because OpenAI just had their dev day. [07:03] some would call this like a Sherlocking event. That's what it's called, right? It feels like every one of their dev days, or every one of their live streams is a Sherlocking event. Oh, I mean, good for them. Yeah, totally. Good for them. But like, I'm really curious how you... [07:16] even can get yourself to write a check and commit to something, because there will be a funding in it. Like, this one, uh, generative AI company, I'm not gonna name it, they had a big, like, announcement the week before, and funding, and blah, blah, blah, and then, [07:30] Sora comes out. 800 million weekly active users, like insane amounts of like power law dynamics there. Like what, like how would the other one win? Yeah. I mean, look, this is, I think this is the thing that every investor is thinking about right now. And I think the best you can really do is invest in great people who are hopefully getting a way ahead of the curve before OpenAI or other incumbents are going to be building something. And in the process, they're building

[08:00] be. I mean, I think [08:02] Yes, it's scary, but I should also point out that this is kind of always how venture has been. There have always been incumbents that can, you know, with one flip of a switch, send a team to go work in your space. I think the difference of what we're seeing here is kind of the ambition [08:16] of Sam Altman and the speed and the kind of ruthlessness, and that's a compliment, to which his organization can move and can move fast. But, you know, when you think about moats, I think what's really important right now [08:29] given how fast OpenAI can move, is getting to a behavior that is sticky and is durable. [08:36] extremely early. And so, you know, to bring it back to something like Granola, you know, Granola is doing incredible meeting notes for anyone that uses the product. Have you used Granola? [08:46] - Mm-mm, nope. - Okay, you gotta use granola. - It's really, really good. - Maybe. Um, so granola does these meeting notes, and it's really, really good. And we find that when somebody uses the product, [08:56] it's incredibly sticky, right? Like the retention is insane. I've never seen retention like this before for a product. Because each time you use it, it just becomes more and more valuable for you. It has more and more of your context and history. [09:08] At a time where all these companies are spending a lot of money for the latest and greatest AI tools, [09:14] What you wanna do, [09:15] is you want to get them to buy as quickly and as early as possible, and get them as retained as early as possible. [09:21] And the hope is that if you can do that, then even if somebody comes along and they build a better granola, you're not going to switch. [09:28] because it has all of your context. And so I think when you hear investors talk about like moats and data moats and things like that,

[09:35] It's that kind of thing that we're talking about. [09:38] Okay, so is that the primary moat you look for? Like, how are you evaluating these companies? So, one of the things that I take a little bit of an issue with right now in venture is, there's so much of a focus on this ARR number, which-- Oh my God, it's fake. Yeah, and it's not even defined properly half the time, right? Like, you might just take-- some of these companies are literally just taking, like, yesterday's subscription rate and, like, multiplying it by 365, whatever. [10:06] I think it doesn't say anything about the durability or the quality of the revenue. One of my partners says this thing that there's like, [10:14] high caloric revenue and low caloric revenue. And it's like you want you want the healthy revenue, right? So I think a lot of rounds are getting done right now based on this revenue. But what I'm really looking at and what all my partners are really looking at is the same stuff we've always looked at, which is retention. You know, how sticky is the product? How often are people coming back in addition to retention? Is it growing organically? Is it growing word of mouth? Or have they just figured out a way to go viral on TikTok a few times? [10:44] If you do that, if you can figure out how to go viral on TikTok, and you have a very well-optimized conversion funnel, you can quickly run up your revenue rate. Like, it's actually not that hard. [10:55] That's how good TikTok is. [10:57] But it doesn't really say anything about the durability or the quality of the product or business. [11:01] For me, it all comes back to the fundamentals. [11:04] is the product retaining users,

[11:06] Is it growing organically? Like, those are the two most important questions in my mind. And when you're looking at the early stages, what kind of check sizes are you writing? Yeah, so this is another thing that's changing as a result of AI. I think, you know, when we're doing really early stage inception type investing, [11:24] It could be anywhere from a couple million dollars, maybe, you know, two million dollars to, in some cases, some founders are raising, call it five or six. [11:32] I think when you're getting more into like a proper seed, [11:35] You know, it's anywhere from [11:37] for or maybe on the higher end, we're seeing some of these larger seeds coming at like 10, again, as a result of AI and the momentum. [11:46] Then Series A's are also getting bigger. A Series A, when I built my company, we raised a $10 million Series A and it was like a big round. It was like, "Wow, that's a big Series A. That can last us a long time." [11:59] you know, macroscope, a series A that I led a few months ago, [12:03] we led that series A with a $25 million check. [12:06] They're getting bigger. They're getting bigger. Carta just put out recent data in their fund performance report, and Series A's... [12:15] are getting a 30% premium. - Wow, what do they attribute that to? [12:20] AI. So, [12:21] AI companies get a 30% premium versus non-AI companies. - Okay. - And it's just like, the biggest difference in this new cycle. Um, and I would say, [12:30] You can imagine that power law dynamics will seek in there, and the delta is probably much greater between like the good teams and like, you know, the average teams. Because once you get into competitive land and you're going after, let's say, like another like huge megafund and there's competition, like you're going to run the price up. Totally. Yeah, that's actually the dynamic we're seeing a lot. And definitely one of the things that's making us want to go earlier and earlier. Some of these companies, you know, they're growing revenue so quickly.

[12:59] that the rounds are happening in such quick succession, [13:02] Oftentimes, if you're the seed investor and things are going well, you might quickly do the Series A. Then by the time a new investor has an opportunity to come in, it's the Series B. It's priced really expensively, but nothing has been de-risked yet. You're paying this really premium price, you're having to write this gigantic check, and you don't know anything more about the business, whether or not it'll fail or succeed. [13:27] I think the natural fallout of that is you have to go earlier and earlier. And, um, [13:33] I think to do that is a new muscle for a lot of big funds. And I also think it's why you're seeing more and more of these accelerators pop up and do so well. [14:03] help you spend smarter and move faster. Their all-in-one solution combines checking, treasury, and FDIC protection into one powerful account. You can send and receive money globally at lightning speeds, get 20 times the standard FDIC coverage through their partner banks, and even high yield from day one. With same day and even same hour liquidity, access your funds

[14:33] and plaid, trust and use Brex. Start today at brex.com slash sorcery. That's B-R-E-X dot com slash sorcery. In today's high-speed business world, staying ahead means using the smartest tools possible, including the powerful capabilities of artificial intelligence. Meet Turing Intelligence. Turing builds customizable AI systems designed to solve your mission-critical challenges, no matter your industry. From expert guidance to tailored projects, [15:03] that's more capable, more adaptable, and more effective. With Turing, discover how AI can accelerate your business growth. To learn more, visit Turing.com slash sorcery. Spelt S-O-U-R-C-E-R-Y. That's Turing.com slash sorcery. Are you concerned at all with that rapid deployment? Going back to this report, like... [15:24] There was a slump in VC, but now we're seeing dry powder deplete within these funds. Like, people are really allocating capital in FAST because it's so competitive. Are you concerned at all with that? [15:54] per year as a partner as we would a few years ago. And I think the only way you can do that is if you hold a really, really high bar. [16:02] And again,

[16:03] don't get caught up in some of these kind of vanity metrics that we talked about, and you go back to the fundamentals. So I think it's really just about [16:11] keeping the bar high. [16:12] Yeah, I want to go back to your point on moats, because I'd also read some tweets, like you've written about moats a bunch, and one of them was on creativity. How do you see creativity as a moat? Yeah, the way I think about creativity as a moat, [16:27] So I've written a lot about creativity in terms of making stuff and tools and... [16:33] art and music and that form of creativity. I think when I talk about creativity as a mode, I'm talking about [16:39] creativity in a business sense, right? If you look at [16:42] every business, every great company, right? Whether it's Apple or Nike or Amazon, and you really study the story of these founders and how these companies came to be, there's something really creative about what they did. There's an ingenuity. It's not something that can just be written by ChachiBT in a, you know, in like an AI-generated business plan. And so, [17:06] I think as it becomes easier and easier to make businesses, which it will because of AI, right? We keep hearing about [17:13] You know, the one-person, billion-dollar company. I'm sure, you know, everyone tweets about that. Sure, it's gonna be easier and easier to execute on business. But I think the thing that will remain rare and special is creativity in business, right? Who is the person or company or set of people that is coming up with something really, really new and novel

[17:34] in terms of a business model or an approach to go to market, or an approach to, you know, how they do their product development. I think the best companies are still going to have that quality. And so, at the end of the day, maybe the only thing that's durable at this point is creativity when everything else can be automated. [17:52] What's a lesson you've learned from one of your founders on this? [17:55] So one lesson for me is actually Suno. [17:59] I think something really creative and genius about what they've done, they have a ton of subscribers and [18:05] a shockingly high amount of revenue that I will let them reveal on your podcast. I will do it. But... [18:12] You know, I think when you're looking at that business early on, what you're saying to yourself is, okay, they really need to [18:18] mirror the consumption businesses of the past, like a Spotify, like a YouTube, like a TikTok. They need to get a bunch of people creating on one side, and then they need to get a bunch of people creating on the other side. [18:32] And to be clear, that is happening. But with the founders of this company, Mikey, [18:36] and Martin and Georg, a belief they had very, very early on, which I would characterize as super creative to both believe this and then pursue it, was that actually in the future, [18:47] AI is going to make it so easy to create and such a personal experience to create. [18:53] that these people that are creating, they're going to listen to their own music. [18:56] and that's a business right there. We're going to get these other consumers over here, and we're going to monetize them just like Spotify does. [19:04] But we think there's also an opportunity to monetize consumption behaviors over here.

[19:09] and not just for creation. And I think that early on, [19:13] I was skeptical. Some of us around the table were skeptical. [19:16] But it was a really creative and sort of bold bet on the future, and a creative business model to pursue. And the thesis has been extremely validated. And I think you only get there by being creative and having the courage to go do it. I've had so much fun using Ceno. It's a blast. It's so much fun. In the world of generative AI, I want to go into Sora. This is like a very hot topic. I'm sure you have a lot of opinions. I have some too. But I want to hear your stance on this. What did you think about the Sora launch? [19:46] Yeah, I thought it was incredible. I thought, first of all, they did a really good job on the product. Just even the onboarding experience of scanning your face was really seamless and fast. And I think they're able to discern a very, very high quality of you just from a very, very brief scan. [20:16] if you remember it, there's a screen that says, [20:19] everything you're about to experience is not real or something like that. It says, you know, like the feed, nothing in the feed is real. [20:26] and [20:27] My guess is they introduced that as a sort of disclaimer, right? Like, hey, reminder, that's not your friend actually stealing that TV from Walmart. It's not real. [20:38] But to me, it almost came off like a mission statement. Like, this feed, you're stepping into an alternate universe. You're stepping into another galaxy. You know, it's-- Ben Thompson, I think, made an interesting comment about Meta's Vibes app.

[20:54] where he said something along the lines of, "This feels like Meta's first real step into the metaverse." [21:00] It's a completely different universe, right? Nothing is real, but it's people we know and we identify with or we may even know personally. So... [21:09] I think that alone just kind of blew me away. And then I also think... [21:14] obviously the Cameos thing was just incredibly smart for a lot of different reasons. I think it's the first time we've seen a new type of social dynamic built on AI. We've seen a lot of these platforms out there. [21:26] try to make the new AI-first social network where they just basically replace humans with [21:31] AI bots or avatars, and I think [21:33] that's boring and it doesn't really tap into kind of some of the [21:38] new and native and interesting things that you can only do with AI, but avatars is like, it's a purely new social dynamic. So, [21:47] I think that's really interesting. I think the biggest takeaway for me, though, is [21:51] what it means for media and the business of media and the economics of media [21:57] Do you remember a few years ago when [21:59] Adam Masseri, CEO of Instagram, [22:02] put out this video and he said, [22:05] Starting now, we're going to be leaning into recommendations. We're going to move away from your friend graph, and we're going to start doing recommendations. Everyone was up in arms. [22:15] Um, all the users, all, especially the influencers. I remember Kylie Jenner put out this huge thing about how she was so upset and you could, you could understand it, right? Because she had something like 900 million followers. It was a business, right? If she could put out a piece of content, tap a button and it reached 900 million followers, that is immense, uh, pricing power and the ability to advertise. So.

[22:40] For her to learn, actually, those 900 million followers, they matter a little bit less. [22:45] that's a big hit to her business. And so [22:47] I wrote this piece, which I call "The End of Social Media," and it was all about this new form of media, [22:53] which I had called recommendation media. And it was really just kind of the TikTok algorithm, and how the TikTok algorithm was sort of the next phase of media. The follower graph was going away, [23:04] And now the platforms were going to get to decide [23:07] what somebody saw. [23:09] far more efficient way of content distribution. I think Sora is kind of the next phase. It's kind of, you know, I hate to say it, it's almost like the end of the creator. Like, if you play this out and think about where this goes, platforms want to amass, um, [23:25] a huge catalog of content, [23:27] And then they want to program each piece of content such that when you're in the platform and you're swiping, your attention is captured for as long as possible. And that's why the TikTok algorithm is so powerful. But it still requires that human beings make the content. And there's a cost to that. And so [23:45] I think where we're inevitably headed is, [23:48] Now when you're swiping TikTok or Reels or Sora, we're eventually going to end up in a place where most of that content is generated on the fly. [23:56] specifically to target you and to capture your attention as long as possible. And unfortunately, like the individual creator just becomes far, far, far less valuable in that dynamic. And so to me, that is the that is the biggest impact of Sora. It marks the end.

[24:12] of the creator, unfortunately, which obviously I find devastating in many ways. Um, you know, I spent, [24:18] my career the first part of my career building tools for creators and um it's it's like it's a whole new chapter for the internet [24:24] Yeah, I think I read a couple posts on different creators saying, like, this is a cataclysmic event. Like, you should, like, really come up with a good strategy going forward. Because it is a doomer standpoint for, like, creators for sure. But, like, I think, like, with all of these things, you do need reinvention. I was talking to PJ, who creates all those, like, extremely viral AI videos. They did one for Kalshi. It reached, like, over 100 million people. Like, it's truly insane. [24:54] And when he did that, I was like, "Okay, I gotta learn how to do that." And this was when, like, Gemini's model was taking off. And so I, like, spent the weekend, like, "Okay, I'm gonna learn how to do this." Like, you kind of have to, like, mix-match on all these different platforms and make something in the end. It was, like, the-- Okay, defensibility, for sure, there. Like, I did not want to do it at the end of the day. Like-- And then with Sora coming out, I DM'd him, and I was like, "Hey, [25:20] What do you think? Like, are you still, like, hiring and training people to, like, do this and create these viral videos for your huge-- by the way, like, corporate, um, customers? [25:32] He was like, no, the game's changed. It's completely changed. I'm only hiring comedians. Wow. I'm only hiring comedy writers. Yeah, I think this is really smart. So, you know, I mentioned that this is obviously really bad for creators, but I think the flip side is it will probably reward uniqueness. Going back to your point or your question to me about creativity as a moat for business, I think now what you're going to see is creativity as a moat for people.

[25:59] Creators, creativity for creativity, right? [26:03] In fact, I think Sora, the Avatar model, is a perfect example of this. Sam actually even hinted to this in his blog post, sort of hinted that there's going to be a new model where they're compensating creators for their likeness, their name and likeness. That's cool. Yeah, maybe we move from a world in which... [26:19] We're compensating people for however many views they can get because now [26:24] Views for a platform are unlimited. They can just get as many views as they want. [26:28] But what they can't get is uniqueness, right? And so if you're a person or a brand or an avatar, [26:34] and you can do something or appear a certain way that is truly unique, [26:39] Maybe that's what's going to be compensated. And [26:42] It's scary in a way to think that the incentives are shifting for creators, but maybe, I don't know, maybe there's something [26:49] Maybe there's a positive sort of silver lining to this as well. Well, I was excited about that because I don't want to make talking head videos. I don't care, but I know that's like one of the viral hacks for like at least women on Instagram. Like it just works. A couple points on the avatars and like the cameos point is I thought it was absolutely genius for Sam to make him one of the stamped. [27:10] Because guess what? Now he's gonna have more videos on the internet of himself than anyone else. Like, the distribution on that is gonna be truly insane. And then going back to, like, understanding monetization on that, I was thinking, "Okay, like, I don't know if I wanna make my cameo public because [27:28] "What are people getting-- I don't want them to misuse this, -or make me do something crazy." - Yeah. - And I was generating-- - But you might for a price. Maybe for a price. You might. Not saying you-- I'm just saying that's probably the calculation -they're making right now, right? - Yeah. It could be. And I was thinking about his because I was trying to recreate-- I had this viral clip of me pretending I didn't know Jack and Sam were related, so I wanted to make a clip of me talking to--

[27:53] Sam and being like, [27:55] The results are in. You're related to Jack Altman. But the problem I had was like, every time I did that, I was prompting it to make Sam like anxious and like scared and worried, and then happy. But like, he was only happy the whole time. Oh, 'cause it's only trained on. What did you engineer in here that you're only happy? That's funny. Yeah. Yeah, I mean, look, um... [28:18] I think people are gonna find ways to capitalize on this. There was a tweet going around today or yesterday about Jake Paul. In six days or seven days, has generated a billion views of his face. [28:32] by doing the same thing, by, you know, making his cameo public, and then I guess tweeting that people should go use it. And so, [28:40] There's going to be a business around this for name and likeness. [28:45] you know, again, at the end of the day, I guess my hope would be that [28:48] Maybe, if nothing else, we can reward creativity and uniqueness and [28:53] people doing something different. So as you said, we don't all just have a bunch of talking head clips floating around because that seems to be the hack that works today. Some of you may not have heard this yet, but our sponsor Public just launched something called Generated Assets, and it brings AI into investing in a way I've honestly never seen before. Here's how it works. You type in an idea like AI-powered supply chain companies with positive free cash flow,

[29:23] then dispatches a swarm of agents that scan every single US stock, evaluates them, and instantly builds a custom index around your thesis. What really stands out is how clearly it explains why each stock is included. And before you invest, you can even backtest your idea against the S&P 500, so you're making decisions with real context, not just guessing. And beyond generated assets, Public lets you invest in stocks, bonds, options, crypto, all in one place. They'll even give you an [29:53] If you want to build a portfolio that actually reflects your thesis, visit public.com slash sorcery. [29:59] Paid for by Public Investing. Full disclosures in the description. [30:03] Maybe this actually just cannibalizes the Cameo. Cameo business is business. Yeah, maybe. Because do you need Cameo anymore if you have this? -I don't know. -Right, yeah, exactly. Yes, that's true. [30:14] - Yeah. - Yeah, so it's really fascinating. I'm interested to see where it goes. So I have an interesting question on Lightspeed's strategy, investing into AI, because you guys have invested into like several LLM competitors. [30:28] What informs your decision there? And I can go through the list unless you have them on top of your head, but you've invested into XAI, Anthropic, Mistral, Stability AI. You've also done Pika and some more. So I'm curious, how do you think about investing in these direct competitors? Yeah, I think Lightspeed from very early on in sort of the AI era, [30:49] kind of revolution took this position that actually AI and foundation models are more akin to

[30:57] software and code than they are to business lines and strategies. And, you know, I think now it would, it would be silly to say X random software company is competitive with Y unrelated software company. It's all just code. And I think that was the, the position we took internally early on. And so as we made these different investments, the way we thought about them was actually, we believe these things are actually quite different, right? XAI has decidedly taken a very direct step into things like [31:27] the X platform and potentially embodied AI and robotics through potential partnerships with companies like Tesla, et cetera, and Optimus and things like that. Whereas Anthropic is like, [31:38] We're very, very focused on the enterprise. We're very focused on code, right? [31:42] And where Mistral is like we're very focused on [31:45] European sovereignty and open source. And so I actually think we view these businesses as quite different. [31:53] And we don't just view kind of AI as being competitive with AI, just in the same way as we don't view software being competitive with software. Yeah, that makes total sense. I think now we're seeing a lot more of the bifurcation between all of them. Exactly. Yeah. And I think that was, you know, when we talked to these companies early on, that's how they were talking about it, too. It's like, this is our mission, this is our vision, we view a world in which we do this part of the world and this company does this part of the world. [32:23] Maybe on the surface early on, it may have looked like from the outside, maybe it looked like that was competitive, but we didn't believe that early on and we don't think the founders did either. Where do you think...

[32:34] the value accrues in the AI stack. [32:37] I think it's such a hard question. I feel like this is the question everyone has. Where's the value accrued? I mean, clearly it's accruing right now. [32:45] to the model layer for companies like Anthropic and OpenAI and X. But there are also amazing application layer businesses generating a lot of value as well. I mean, we're not in this company, but there are many reports out there about how much Cursor is generating at the application layer. And there are lots of other companies that are generating an accruing value at the application layer. And so, [33:10] I know it's a cop-out answer, but the simple answer is, it's at whatever layer somebody is willing to spend money. If you're a huge foundation model and you can hit every layer of the stack, [33:22] there's probably more revenue for you to take advantage of than if you're just at the application layer. [33:29] We're seeing now that... [33:30] Every different layer of the stack is able to [33:34] generate an interesting type of business. What about your AI education startup? [33:39] Yes, Oboe. My anchor co-founder, Nir Zuckerman, for a long time has been thinking about [33:49] education and about how [33:52] We as humans, we are spending however many billions of dollars on making machines extremely intelligent. But one thing that we're not doing is we're not. [34:03] doing the same thing for human intelligence. Mm-hmm. And so he and I have a belief that AI can actually be used to make human beings smarter. And people are obviously using AI and Foundation models to teach themselves.

[34:16] But what no one has really done yet is focused on the products, right? Like, how do we make great products to help people learn? Everyone talks about, "Hey, it'll be amazing one day when you have a tutor that knows everything about you, knows everything you've ever learned, knows everything about how you like to learn, what the most efficient way is to teach you, and then actually teaches you things that makes you smarter." [34:38] And so that's what we're thinking about with Oboe. We want to build the products that leverage AI to actually teach people things. And so we launched the first product a few weeks ago. It's called Oboe. The whole idea of it is that it creates [34:52] these magical educational courses for you on the fly about anything. [34:57] and in a way that's customized and personalized to you and what you like to learn. And it enables you to learn them in whatever format best suits, again, the way that you like to learn. So we can teach you through audio, we can teach you through text. That's the basic premise of the company. And, um... [35:13] It's been really exciting to work again with my anchor co-founder. He's taken the lead on this one. He's CEO. I'm just riding his coattails as a glorified advisor and board member and co-founder. [35:28] Education has always been like a really, really hard space to get returns out of, build companies, get adoption. Why do you think now is the right time? We think about this less as ed tech, and we think about this more as a horizontal need. [35:43] for every person on the planet. We're all learning every day, right? When you go to Google something,

[35:49] you're learning. When you go to ChatGPT, you're learning. But today, what you also do is you go to Google, you end up on some website, [35:56] then you realize you didn't get what you wanted, so you go over to YouTube, you search for a video, you watch the video, you bounce around the timeline. You're kind of like piecing together the educational journey. And yeah, you might come out of it a little bit smarter, [36:09] but you wasted a bunch of time and it didn't actually teach you in the way that's best for you. [36:15] But with AI, we can effectively generate that entire experience for you in real time. [36:20] based on anything we've already learned about what you like to learn or what's effective for teaching you. And so we don't view it as like, "Hey, we're going to reinvent the school." We view it as more every single person on the planet [36:33] can be learning far more efficiently than what the horizontal tools offer them today. And that's the hole we're going to try to fill. [36:38] What do you think the world's going to look like once we have superintelligence? Uh, well... [36:44] I think the key is to [36:46] I'm going to channel my inner Ilya Sutskavar here, but I think the key is aligning intelligence with the needs of humans. And I think that that's a really tough thing to do. [36:59] uh fortunately we have a lot of great minds thinking about it but i think [37:03] superintelligence, [37:06] is going to be able to do things in a far more capable way than we're going to be able to do and think, [37:12] however many steps ahead of us. And so the most important thing is making sure that we can be aligned [37:18] with it from the beginning?

[37:20] Founders ship faster on Deel. Set up payroll for any country in minutes, hire anyone anywhere, and get visas handled fast so you stay focused on scaling. Deel takes care of onboarding, HR, IT, ER, benefits, and compliance so your team can grow without borders. It's why more than 37,000 fast-growing companies trust Deel to move fast. Visit deel.com slash sorcery. That's deel.com slash s-o-u-r-c-e-r-y. [37:47] Coming from the podcast world, what-- How big do you think this will get? - Podcasting? - Yeah. Yeah, so I've-- I've-- I've frequently underestimated, I think, how big podcasts can get. We thought it was this niche little thing. [38:00] Uh, then we saw that it was becoming a big thing, and all of the biggest companies in the world wanted a piece of it. [38:07] And then, you know, I think around the time that I left Spotify in 2022, I [38:12] I sort of personally felt like, okay, [38:15] The work is done and it's great and I'm happy with where it is. [38:18] And then I think once again I was proven wrong because in 2024, 2025, as we saw, we had the podcast election, right? And now I just think that podcasts are basically the new television, right? Yeah. In the same way that you and I grew up watching TV, I think now people watch podcasts. And [38:37] Yeah, today, most of them are maybe people sitting around the table having a conversation like this. But remember that just a few short years ago, podcasts weren't even associated with video. [38:48] Video is a new format for podcasts. And so I think as people get more and more used to video being the medium, people are going to be more and more creative and fill the space in new ways. I mean, look at your show. You're doing it from all these different, interesting locations. You're not just sitting behind a mic at your computer and using, you know, an amazing tool like Riverside. You're like out there in the real world and

[39:10] I think we're going to see more and more people push [39:14] the boundaries of the format. I think when it's all said and done, again, I think it's-- I think it's television. It's-- It's television that anyone [39:21] can broadcast, right? And I think if you view it through that lens, we've got a long ways to go before it reaches its peak. From working at Spotify, what is the biggest lesson you learned from Daniel Ek? [39:33] I think one of the amazing things about Spotify and Daniel is there's really like no ceiling to the ambition. [39:41] He started Spotify in a world where, because of piracy, music was effectively free. It was illegal to be free, but it was effectively free. And he built... [39:50] a platform [39:52] and a company [39:53] that was effectively better than free. [39:55] and that was the scope of the ambition. And you've obviously seen that with some of the other companies. He started companies like Neko, which Lightspeed is fortunate to be an investor in. But also with his business strategy, [40:11] and ambition inside of these companies. So Spotify... [40:15] I think really smartly, [40:16] took their position in music, built up over the course of, you know, 10, 12 years, and said, "We're going to take this position of strength where we have, you know, [40:26] hundreds of millions of people listening to music, and now we're going to ladder into a new format that sits adjacent to music and has all the same interface, right? Like, if you think about music, [40:39] it's largely the same interface as podcasts. So we can ladder into podcasts, right?

[40:44] And then once you do that and you do it successfully, it's like, okay, what's the next thing we can ladder in that leverages off everything else we've built? So they go into audiobooks. And so I think the biggest thing that I've taken from that experience was just, [40:57] how smart and strategic you can be as a big company when you've sort of like won a position in something to be able to leverage that to go into the next thing. That's been a, that was a really, really big learning from being there. And I think [41:09] Spotify in general is a very, very strategic company. They take [41:13] a long time to make decisions. I don't mean that in a bad way. I think I know every company wants to be fast, but [41:19] You know, they're thinking about the future, [41:21] five ten years in advance and i think most of us don't have the attention span to do that [41:26] Yeah, that's a really good segue into my Brex question. [41:30] So, Brex is all about performance. [41:33] Spending smarter, moving faster, the modern intelligence finance platform. But I'm curious from your standpoint, like, how do you view performance as an investor as you handle all of these portfolio companies? Like, when to lean in, when to, you know, let someone bake a little bit more. How are you thinking about that in terms of like, okay, like... [41:52] Talk about the Spotify strategy of, like, turning the dials at certain times. Yeah. I think of-- especially because I'm so focused on early stage investment these days. You know, I think performance can easily be measured by the standard things: revenue, Dow, now. But the way I like to think about it, the earliest stages is [42:11] speed of iteration, speed of shipping, speed of learning. Like, how quickly are you getting stuff out there into the world, and getting signal back, and learning, and then making the next decision, just a little bit smarter? This has always been a thing I've cared about, and it's obviously always been important for startups. But again, with AI, everything is moving so fast, you have to be willing to go faster, and you have to be willing to take more and more risks. So I think about performance as speed. And so then as we close out, I love to end on predictions.

[42:40] So which companies do you think will officially announce an IPO this year? Give me like three. [42:47] which companies will go public. [42:50] Boring answer, but open AI. [42:52] You think OpenAI is going to go public this year? [42:55] I'm in. [42:56] They're the biggest private company, right? [42:58] I'll throw out another one. [43:01] just for fun. [43:03] Calci, I saw they raised a big round, so maybe they'll go public next year too. [43:07] Polymarket, [43:08] I know a competitor, but also raised a very big round, partnered with New York Stock Exchange, ICE. Maybe they'll go public. [43:14] Cerebris seems like a candidate. I know they were going to go public. I think they actually just like withdrew their plans to go public, or at least I read that. So that seems like another one that could happen. Discord, I feel like has been baking for a really long time, large successful platform. [43:30] That seems like a good candidate. And then, yeah, I mean, I feel like any of the AI incumbents or hyperscalers like OpenAI seems like they have a great shot as well. Do you think that might be the next strategic, like, [43:42] momentum milestone. I feel like they're always trying to outpace each other with different kinds of partnerships and announcements and growth and blah, blah, blah. Do you think-- [43:52] IPOs might be the next layer? Maybe. I mean, one of the things that I've been thinking about, a friend of mine, really smart friend, [43:58] said, "You know what? Maybe [44:00] maybe more AI companies should be going public because there's clearly a lot of demand for it. And if you're a public market, you know, if you're a public investor, a public market investor, a retail investor, you don't really have a lot of choices, right? You've got, you've got, you've got,

[44:14] Google, you've got Meta, you've got Nvidia, [44:18] And and so maybe [44:20] some AI companies maybe who are even earlier on their trajectory and maybe have less revenue than we've been seeing from typical companies that are going public. Maybe they should consider going public because there's probably just so much [44:33] pent up demand from retail investors. So I thought that was an interesting take. That is really interesting. [44:39] Okay. Well, Michael, it was a pleasure to have you on. Pleasure is all mine. Thank you so much, Molly. [44:44] Thank you.