Moritz Baier-Lentz on why the future of AI runs on video games

Moritz Baier-Lentz is one of the foremost investors in gaming, AI, and frontier technologies—including companies like OpenAI, SpaceX, Palantir, Etched, Decart, and Recursive. He founded Goldman Sachs's global gaming practice and was previously a partner, IC member, and head of gaming at Lightspeed. Moritz is part of the founding team of General Intuition, an AI research lab training agents on video game data, which has raised a $133M Seed round and a $320M Series A. He is also a senior advisor to OpenAI, McKinsey & Co., and TPG Capital.

In this episode of Summation, Moritz and Auren discuss:

• Why video games are the best data set for the next phase of AI
• The idea that humans only truly learn during play
• Why Germany's fear of risk is cultural
• Getting into Stanford with an essay about Diablo II

You can find Auren Hoffman on X at @auren and Moritz Baier-Lentz on LinkedIn or his website.

Transcript

SPEAKER_02 0:00

I think there's only two things in life. There's work and play. I find it shocking instead how poorly AI is received in the gaming industries. I raised over two billion dollars this year with my portfolio companies. Half of that was done without a pitch deck.

SPEAKER_00 0:18

Hello, fellow data nerds. My guest today. My guest today is Mortz Bear Lentz. Mortz is one of the foremost investors in the gaming world. He founded Goldman Sachs's global gaming practice and is advised on gazillion dollars of transactions over the years. He's also a founding member of General Intuition, which is a research lab training AI agents on video game data. It raised $133 million seed funding. And I just found out that you just raised another $320 million Series A, which is pretty awesome. Um, he recently left Lightspeed Ventures, where he was a partner and head of gaming. He also serves as a senior advisor to McKinsey, OpenAI, TPG, and does a whole bunch of other stuff. Mortz, welcome to summation. Thank you. Thank you. All right, before we get into it, okay, like the first question is you have like, I don't know, 400 jobs. Like, how do you manage all this stuff? Like, walk us through like a life in Mort.

SPEAKER_02 1:15

Well, slightly less than 400, luckily. Um, but it's probably about 15 or so uh right now. And um, you know, since April, after leaving Lightspeed, I spend the majority of my time, I would call it about 80%, um, you know, investing my own capital now and then advising the founders that I work with quite closely. I would say that most portfolio companies and most founders I invest in, I'm really in touch with daily. Um, and it eps and flows where, like in certain months, I think with some portfolio companies, I've spent as much as a hundred hours in in a given month. Now, that typically is during fundraising um or kind of like the PR and marketing push that comes after it. So there's certainly like peak times and downtimes.

SPEAKER_00 2:02

Are you using I and I know you and I have had many discussions about how to use AI? Like, how are you using AI is kind of your second brain like to actually make all this happen? Because at some point you do have to sleep, right? That's true. Every Wednesday.

SPEAKER_01 2:15

Um yeah.

SPEAKER_02 2:16

I mean, just to just to finish the first question, the the the investing, the solo investing and advising founders specifically and startups, that's really 80% of my time. And then the other 20% are uh are those senior advisor roles that you called out, where um between OpenAI, McKinsey, and TPG, I helped those three firms specifically with their foray into gaming and interactive media. And I think also increasingly, really, with the AI transformation of the gaming and interactive media industries. So that's my jobs in a nutshell. Um, AI efficiency. I would say I'm a power user of uh intelligence. I'm not a power user of agentic uh use cases, at least for the time being. And we've spoken about that. I I personally think for anything outside of uh coding applications, because the domain is just so much less verifiable, it's not quite there. Um, but what I'm doing religiously, as you and I also have discussed, is I'm really trying to record as much as possible. Um, my thesis is that uh once agents are good enough to genuinely and in high quality support all the things I do on a regular basis, which excludes coding, um, but includes a lot of other systematic tasks that probably will get there over the course of the next maybe call it 12 months. Um, right now I think the best thing to do is create as good of a context window of my own work and what I consider success for these activities. And that includes, you know, recording uh obviously uh like digital conversations as they happen through Zoom. It also includes uh recording conversations in in my personal life. I I use a range of uh tools and and and you know on Zoom I really like granola. Um there's uh there's a um hardware solution for phones uh called Plot. Um and then you know there's there's so much context in in many other applications. So we talked about Google Takeout and how it's probably helpful to really have this massive, you know, markdown repository of um of digital activity.

SPEAKER_00 4:27

You're taking like Google Takeout, which is like your search history and a bunch of other kind of like interesting things, where you went and stuff like that, and you're bringing it into like an obsidian or something.

SPEAKER_02 4:37

Yeah, I think obsidian is kind of like the yeah, the choice. And and again, this is really right now, it's a massive data repository. Uh the the key will then obviously be connecting that to capable agents. And and again, just to make that point, at least my personal opinion is that they're not quite there yet. But I also think they will be very soon.

SPEAKER_00 4:56

Now, if you're doing something just like online that you don't think your agent can do for you yet, like I don't know, buy some shoes or something. Are you recording the video of like you doing that so that you can have that in the future so that the agent understands your tastes and your preference? Or are you not actually are you recording like your browser use and things?

SPEAKER_02 5:16

No, but I mean obviously there's browser history, right? So that's part of the the Google takeout uh universe is is is your Chrome uh history, and and I think you know that'll be a great representation of your shopping behavior and taste, etc. And what about like emails? Are you connecting that into all of them? Yeah. All emails, all calendar events, all all physical movement on Google Maps.

SPEAKER_00 5:39

Everyone has like this kind of gradation of, okay, I'm I'm willing to give this to the I'm willing to give this to AI and whatever, whatever it is, Claude or OpenAI or whatever it is. I'm not willing to give this other thing. Um sounds like you're you're a little bit more on the because you're willing to give like all your emails and stuff there. Like, where do you is there certain things you're like, okay, I can't give that because there's a security problem because that's where I get my password resets or you know, like how how do you think about all that?

SPEAKER_02 6:07

I mean, passwords live outside of the data that is that is fled. Um as far as I'm concerned, I'm not engaging in any criminal activity. And so maybe um others that are more hesitant have have a different answer to that question.

SPEAKER_00 6:20

Okay, got it. So you're just like, let's let's go all in because it's a great training set, let's get as much training set as possible.

SPEAKER_02 6:28

So in that way, like for now, all of this lives locally, right? So so part of my decision ultimately will also be uh when this is when this is ripe and ready for agentic solutions, where this goes and what of it goes into the agentic system.

SPEAKER_00 6:42

Now, today I assume you're already asking, you've got a big obsidian file, you're already asking questions of that obsidian file. I do something similar where I'll like ask questions just of my obsidian file. Like what types of questions are you asking of that file?

SPEAKER_02 6:56

I I don't I don't actually today. Oh, you don't? Okay, no. Okay, no. This is this is this this lives completely. Just like for the future for the time being. Okay. Yeah.

SPEAKER_00 7:04

Okay. I'll do something like I'll say give me a restaurant recommendation based on all the I've got 14 years of history restaurants, like based on what I'm doing, I'm gonna be in Denver, give me a restaurant and history, you know, that that type of thing.

SPEAKER_02 7:15

I mean, I've found even for that kind of stuff, just just just using the plain tools. I think they have enough context just from the sum of interactions with them that they give me quite reasonable answers. Yeah.

SPEAKER_00 7:25

Okay, cool. Um, now you said the next you know, billion dollar games might not look like games at all. Like what will they look like?

SPEAKER_02 7:34

Yeah, I mean, just to kind of take a step back, there's been this huge uh interest in um VC funding and gaming, kind of starting around 2019-2020, and then certainly exacerbated by the the COVID and metaverse ambitions through 2020 and 21. I think this has faded quite a bit uh over the last two years. But from a venture perspective, it's probably always true that um it was maybe less interesting to fund individual games and and experiences. That's more akin to kind of movie financing. I I would say that's not really in the domain of generating, you know, what ultimately turns the needle for a scaled VC fund, which is like a you know 10, 20, 50, 100 billion dollar plus outcome at this point. And so the the question becomes where can games create scaled consumer experiences that maybe resemble social networks? I think Roblox is a great example. I think Discord is a good example. So those were, I think, the I mean is TikTok a game, I mean, in some ways, is you know, it's a game.

SPEAKER_00 8:42

I mean kind of. Yeah. It seems like it's a game, right?

SPEAKER_02 8:45

Yeah. Well, you can, I think my from a philosophical perspective, when we can get into these things when we talk about general intuition, I think there's only two things in life: there's work and play. And so um, and I think you only learn during play, and we can unpack that as well. But the more interesting stuff, even about games, and I think where I'm focusing my attention now is you know, there's this narrative that gaming is a $200 billion industry and it's the largest media category, and so on, and all of that is true. But $200 billion, even in the grand scheme of things, is still small. Like it's large enough for an open AI to care about it, but it's not gigantic, especially when compared to the enterprise software industry. And this is also why VC attention isn't necessarily there, at least from like the scaled funds. However, play and concepts of play are massively important to the to the human existence. And um, what we do with general intuition, for example, is this is not uh we're not building AI for games. We're actually using games to advance the state of AI across all sectors. General Intuition is a is a company that um leverages gaming data and leverages gamer performance in virtual worlds, not to build any games at all, but to build intelligence that is more deeply rooted in space and time and basically expand the entire catalog of machine intelligence, which so far has been rather reductive and limited to language and text and symbols, uh, you know, helped by obviously the breakthroughs of the Transformer and the fact that all of this was neatly scrapable online. But if you really think about how human learning unfolds, um, this is very deeply rooted in space and time, like a toddler learning to walk or an animal learning to climb. And so the question in ML, and I did my undergrad in in ML, and it was all about kind of reinforcement learning in space and time back then. And I think we're coming back to that a little bit is can we collect human action data in a wide range of environments and measure the outcomes, but also the action and intent that led to certain activities to begin with? And can we start training these uh foundation models on along those lines? Like world models is a term that's now quite popular, and I would say has has entered the the AI mainstream maybe about a year ago. But it's something we've been fascinated with and working on for much longer than that, and and that laid the groundwork for general intuition, for example, and other companies I'm involved with.

SPEAKER_00 11:17

If you go to like a history of AI all the way back, um, and then you know, even recently with like DeepMind, so much was about like using the AI to play the game, learning from the game. Okay, can we go beat uh grandmaster in chess? Can we go meet the grandmaster in Go? Can we do this? Can we play StarCraft really well? Can we do like so much of these things have been based on game? Why is it just because it is like it's a great kind of confined area for us to learn on?

SPEAKER_02 11:46

Yeah, I mean, first of all, uh also to point it out, Nvidia, the most valuable company in 2026, has been a games company, right? Making, yep, making I I think part of that is um games have uh often and I think still are um on the frontier in terms of technological necessity to make them actually really enjoyable, interactive experiences, and games as a as a category has been pushing the technical boundaries to the extent that sometimes the technologies built for gaming then also found applications outside of it. And I think GPUs are the most phenomenal example of that and and the story of an NVIDIA. Um, but uh then abstracting a little bit play, you know, play, you can make the argument uh almost from an evolutionary perspective, is kind of nature's way of um uh maximizing informational return while keeping the stakes low. So you create almost these alternative options and universes in your head, you act them out, you test the environment, you collect the data, and failure doesn't have severe consequences. Specifically, for example, you know, let's then and this is something we don't see only in humans, right? We see this in animals too. Like, why why are we fascinated by play? Like, why why why is play fun or rewarding, right? Like most of this can can typically be evolutionarily explained. And I think you want to know how to climb before the saber-toothed tiger comes around the corner because you're only trying it then for the first time.

SPEAKER_00 13:34

You're probably gonna have a rough day. It's even like the uh, you know, playing hide and seek when you're two, or I'm gonna tickle you, type of things. Like, even those types of games, like, do you learn a lot from those things?

SPEAKER_02 13:45

Cups messing around with each other, you know. Again, like that we see this in animals, we see this in in in in humans. Um testing boundaries, also, right? Like, I mean, and and sometimes playfully and joyfully. And I think you could go as far to say as just briefly mentioned, you know, there are some theorists that that proclaim that um really learning happens during play and and and not during work. Work as a juxtaposition of kind of taking everything you learned in play. And we're we're talking in abstract ways here, right? When I use these words play and work, it's a little abstract, but but work then being uh taking everything you learned during play and kind of quote unquote putting it on the road and executing and reaping rewards, but not really trying new things and learning new things. And now fascinatingly, um this dichotomy also is represented in um in many you know AI machine learning paradigms where there's always this um subtle trade-off between exploration and exploitation, which you can almost take synonymous with with play and work. And so these concepts, um, yeah, these concepts are very innate to human intelligence, and at least so far they are innate to how we've built machine intelligence.

SPEAKER_00 15:08

Why are sports so popular?

SPEAKER_02 15:10

Like it seems like is that just like the play to the next level? Amateur sports and playing sports, I think, is enjoyable because it is a form of play. You go play tennis with your buddies, you play pick up soccer, you know, whatever it might be. Yeah. And then I think watching sports to a large extent is also kind of like uh, you know, re remnants of uh tribal fantasies. Um but one thing specifically about video games too is like I this this idea that we would train uh foundation models on um on spatial temporal data rather than uh kind of linguistic fragments, and language is ultimately like a lossy abstraction of the real world. I think you know everyone would would agree with that. It's not new to say, hey, let's take human action data in the world. We have a lot of uh we have a lot of world data and environment data of humans. We have plenty of recordings. Uh I mean any any recording of a human in any environment doing anything is a great kind of capture of the um of the results or the environmental response of our innate actions and and intent, but we don't often have this action and intent data, not in a very kind of precise, discrete way. How would we go and and capture everyone's actions that led to the stuff that we see in the world? This is like a little bit almost like kind of behavioral uh psychology, but you would have to plug uh everyone into a brain computer interface and kind of like read the signals before they manifest in the world and then measure the response of the world. We're not there, and we won't be for a while. Um, and so ironically, the best proxy we have for tightly coupled human actions in worlds is the sum of humans playing video games, um, you know, across the the many thousand games. Uh games have other interesting features too. They um they're very discrete, they're very neatly measurable, their states are very clear, the goals are typically very clear, the feedback cycles are uh uh short, and you can do things that are quote unquote out of distribution. Uh playing uh Microsoft Flight Simulator is very different from flying a plane. You probably wouldn't try doing a looping and a Boeing 747, but um, we can figure out if it's possible or not in Microsoft Flight Simulator. And it's a very realistic um and and complex experience. And so ironically, and this is the big insight that led to general intuition, uh, the the sum of uh of coupled action and environment data was accumulated through the the preceding company and platform metal, uh ultimately led to what is, as far as we are concerned, and kind of tested this hypothesis, even with big frontier labs, and including OpenAI and DeepMind, this is the best data set for world and action modeling. And so we're taking this not to then build games with it or build media experiences with it, which you can do. If you have action and world data, you can do two things. You can generate a world if the user gives you new actions. These are like the world models or kind of like playable movies, or like you know, if people have seen um a Deep Minds Genie, for example, was I think eye-opening for many people. You see a screen and you it's almost like a playable video. You press forward and it gets rendered as you go. So you generate a world given an action. However, what you can also do, and now we're talking about servicing a 200 billion TAM or whatever kind of like games and and film is combined, let's call it like 300 billion. Um, however, when you have this data set, you can also do you can also generate actions given worlds. So so you can for any type of embodiment, whether digital or real, uh for you know, bipedal or quadruplets or or whatever, um say, hey, tell me what you are, screen me where you are, and tell me your goal, and I will give you back the optimal action sequence to accomplish this goal. And now then the question is what is the TAM of problem solving and goal seeking in space and time? That is literally that is what we do as as humans, right?

SPEAKER_00 19:30

So what no certain why there are certain folks in like the gaming industry that seem like very AI forward, and certain folks in the gaming industry that seem like quite hostile to AI. Um, and I would think like just based on the people I know, everyone would be pretty AI forward. Like, why is there this hostility?

SPEAKER_02 19:48

Yeah, I think your your sample size is certainly skewed in a in a in a certain uh direction, but you're right. I find it, and I've said this before, and I've literally been quoted in in headlines on this. I find it shocking and sad uh how poorly AI is received in the gaming industries, specifically, I would say, by game developers, um, less so by players, actually, and certainly at this point, not by the executives steering these companies, who are also obviously all aware that this is the future and aren't just like I'm worried about my job type of thing, or is it more is it more visceral, or it's like a craft? I would say, and it's particular against what we just discussed, this backdrop of uh gaming um always adopting and actually pushing and creating new technology. It's it's quite uh yeah, I ironic, I think, that now there's this resistance. But I think it's the following gaming has had a rough three, four last years, and by and large, I'm oversimplifying a little bit, but by and large, COVID was a huge boon for gaming and any kind of inside digital activities. There was a lot of overinvesting, there was a lot of overstaffing, um, budgets ballooned, um, not often followed with uh the blockbuster uh successes needed to kind of warrant those investments. And so we've been we're coming on four years now, I think, of record layoffs in the gaming industry. So that's an important backdrop. And um, and also, you know, this is an industry that takes real pride in their creative uh tasks. And I think uh other tools maybe felt more like they were enabling you, but not infringing on your taste or or or or kind of secret sauce. And this one feels different. And I think when you're in an environment where there's a lot of anxiety, um, there is it's always easier to point to external factors than to maybe, maybe look, look internally. I think it's a very natural reaction. The other thing that's also contributing to this in gaming is unlike because you would you you saw the same thing in Hollywood and and And movies. And I think actually it's turnaround there, and there's no longer a debate in Hollywood and movies whether you should use AI. It's like they all kind of have to. I mean, you can see also like how the industry is involving evolving here in Hollywood, and and you know how many movies are actually still produced in Hollywood. So they get it, but they also have things to point to. You can look at AI video now, and it's really damn good. And I think three years ago everyone was doing the same. They're like, look at Will Smith eating spaghetti, if you remember those videos. It was a hot mess. And now it's beautiful in HD and like high fidelity, and you couldn't distinguish it from reality in certain cases. Um, the gaming industry, interactive is harder, right? We figured everything else out. We did like text, image, audio, video. Is it still all linear? Interactive is really, really, really rough. Um, because you have to really get it right. Because if you mess up even just one percent with every interaction, everything degrades to just slop at zero. Yep. So it is significantly harder. But um because of we're gonna get there, but it's also easy right now, I think, for people to point to like AI games and say, Oh, look, it's and I'm like, guys, just just look at the conversation we had three years ago on on video or two years ago on videos. And so, yes, we're gonna get there. Um, the sad irony is that people are revolting and raising their pitchforks uh in fear of their jobs and in wanting to keep their jobs. Yet, and I've said this many times too, the only the only certain way to lose your job, probably or or increase the like likelihood vastly over the next few years, is to do exactly that and not experiment and adopt these new technologies.

SPEAKER_00 23:42

Not try to use it.

SPEAKER_02 23:43

It will be a reality. AI will significantly change how games are built, not just how they are built, but even it will create previously impossible experiences that you can't even build without AI. There will become a new standard, there will become a new like player and user uh requirement and need that that you can't even service uh if if you're not an organization that has built AI native. And that transition takes time. I'm also I'm not envying incumbents. It's very easy for me as an early stage investor. Right. I'm just gonna throw my money at some cool that with high likelihood won't work, but if it does, it changes everything, right? That that's like world models and action models and like AI and PCs or whatever. Very easy for me to say this and do this. I also don't envy the big operators that need to kind of like segue into a new reality over time and tactfully, uh, working within the constraints of today, but but being ready for quite different constraints in just a few years. So it's that this is the kind of stuff I work on with OpenAI and McKinsey, for example, which is helping the incumbents with this transition in a way that's you know reasonable and and meaningful and practical and and and and and financially responsible. And then the work where I spend my 80%, like where I'm me, me as the startup, investor and contact supporter, will go all in just on the absolutely crazy stuff.

SPEAKER_00 25:08

Perhaps like one of the most successful gaming companies, so not everyone thinks of them as a gaming company, is AppLovin. Why have they been so successful?

SPEAKER_02 25:16

AppLovin uh is actually one of the few scaled examples. I mean, they build a whole um uh ad network, right? And that that's a that's a tough thing to do technologically. Yeah, you have to you have to get the the supply right, but um uh yeah, I mean you can call them an a gaming company or gaming adjacent company. I don't Adam is a friend, he's uh he's a smart guy. It was a long haul to get there. Um, and and you know that really broke out the building the adward network, as far as I'm concerned, wasn't always the strategy. So um, but yeah, uh I think a good example of the the the insight that um uh uh it it's always a good idea to uh seek novel uh novel things and and maybe uh maybe build things that others aren't building and try and reason from first principles. Because I'm sure there were a lot of people that told them you can't just build an ad network. There's uh there's already Google and and Meta, and they're doing quite fine.

SPEAKER_00 26:18

What do you think other businesses can learn from this kind of like free-to-play gaming model?

SPEAKER_02 26:24

Free-to-play hasn't been necessarily like a uh um you know smooth ride and and uh a fairy tale of mostly successes. I think it's it's very, very hard to build free-to-play because you have to assume a certain scale to make it work through through microtransactions. I think um what's there to learn from games like Roblox and and Fortnite specifically is that they have transcended to to our earlier point this notion of becoming games where you know I think many people who who play these wouldn't necessarily even call them gamers today. It's just that their social hangout spaces in the kind of like third spaces. I think what they really amassed is your social graph is there. And and by the way, that's also a big uh it's very difficult right now for new games and new platforms to emerge because you really have to crack through this social graph and attention ceiling to even have a shot at at kind of building a new social graph in your universe and in your experience. So they're becoming sticky. We we see a lot of consolidation right now, uh especially around Roblox, which is something that um uh Tim Sweeney at Epic has also called out. Um in a way, these are becoming you know pretty walled gardens, uh scaled walled gardens that that make innovation in the industry somewhat harder.

SPEAKER_00 27:52

When when you think of these companies, the Minecrafts, the Roblox, the Fortnites, um like why why have these these communities been so successful and other ones which maybe also like looked amazing, seemed really cool, somehow they didn't they didn't hit. Like what what did they do differently? What can we learn from them?

SPEAKER_02 28:14

One thing I kind of love to mention when it comes to to game design or experience design is um you want to make sure that uh that the game allows for players to be fun and makes the players more fun rather than actually be the game and the content being the fun. You'll inevitably end up in some kind of you know content threadmill. If why people are playing this game is because of the the the IP and the lore and the narrative. And you look at kind of like the ones that break out the most. I mean, again, to come back to like these sandboxy experiences like Roblox and and Minecraft, certainly, but even Fortnite is kind of like this this smorgasburg of where people can just go wild and and they just have a good day. They're gonna be a good thing. Yeah, self-expression, social roleplay. Um, those are huge, like even GTA 5. I mean, GTA is a very popular game, but you know, the the most popular mode right now for GTA 5, which is now many years old, is is a mod that they acquired uh from 5M uh that enabled uh GTA to be like a multiplayer social roleplay experience, and people just having a blast, like being a chicken on the road uh to to crash someone else's Ferrari, and then that lands on TikTok, and and and you know, there's like proximity voice chat and and all this kind of stuff.

SPEAKER_00 29:41

So you kind of need both, right? You need like you need the like the goal of the game because I remember like this is like 20 plus years ago, like playing, you know, being in Second Life, um, if you remember Second Life, and it was like, I don't know, it's kind of boring. Like I tried a few times and like seemed cool, but there was like I wasn't sure what I was supposed to do. There was no like goal of it. It is it is kind of important to have both in the game, right?

SPEAKER_02 30:03

Or yeah, I mean this was like the the the this was like the the the empty metaverse before the the bigger empty metaverse. Um yeah, you uh again, like games and online experience must be must be vessels to kind of allow players to be fun or or and engaging. Um yeah, you yeah, you you're probably right. You you want to make sure you feed it with content. There needs to be a reason to come, but then also you it will only go mega scale because of the players and the ingenuity of the players.

SPEAKER_00 30:39

It will never go mega scale because you you infinitely build new, fresh content as a as a publisher or there are people who play a lot of games that could be video games, board games, card games, like that kind of there's those types of people. And then there's all these other people who don't do that. Um, and so some people just like gravitate to playing games in some sort of way, and some people don't. What is is there like a personality trait? And when you're working with people, are you are uh it seems like you're preferring more of the gaming people, or maybe not?

SPEAKER_02 31:11

Yeah, I think on on player motivations, at least as uh as it comes to video games, there are different um uh player motivation types. And and when you design your game, you obviously want to solve around that. There are people who strive for mastery, right? They really care. Like that that was like my whole childhood was was that like the thing I cared the most was trying to be the best in the world and what I was doing. And I had a blast alongside it and I loved the game. But what that was the thing that kept kept me coming back every day. There are uh there are people who like the social element of it, there are people who are the explorer type, uh, or like collecting achievements and and kind of completeness if you think about Pokemon Go, for example. So um there is no one player motivation.

SPEAKER_00 31:57

And some people want more novelty, like they're just like like I know people who play lots of board games, but they they get like they get bored pretty quickly of like a board game and they're always want to move to like the next new board game or something like that. And then I know people who just like are just they just play Settlers of Catan every night, kind of kind of thing, you know. Um and they they look they want to get really good at one thing.

SPEAKER_02 32:16

I've played literally like one game so far this year, which is uh which is Rocket League. And and and and that's now, I don't know, eight years old. So so yeah, so there's no no one fits-it-all motivation. In terms of founders I work with, I mean, I've obviously been focused on the gaming and and interactive media industry, at least for the last 10 years. Um, some of my best investments have been in in Frontier and pure play AI, and now in my my own capacity, um, I work with uh founders from a range of industries. The one thing for me that all of them have in common, or where I gravitate to is um I usually like to work with people who are really pursuing audacious North stars, or let's call it like global optima. I'm not really interested in uh um in something that that doesn't uh scream N of one company or at least kind of like is in that vein. So like someone making another game that's not interesting by definition. Someone building another like enterprise software of this sort or that sort. And and enterprise software in general is is a domain where I don't spend much time. Um, even with AI innovations, I want things that uh no one else, even including the large labs, are doing. General intuition and and world models and action models at the time is a great example. Like this is genuinely uh an advancement of even the catalog of uh foundational AI technologies. Another good example is recursive, a team that's um saying, like, hey, uh, you know, people realize that these models uh don't don't use uh enough space and time data. That's general intuition and Descartes, and we're trying to like improve these models uh based on how human learning really unfolds. Another way in which human learning is today quite advanced compared to what we're building for the machines is continual learning and this idea of recursive self-improvement. Right now, the models are trained. And and of course, like we, you know, the the big labs they use their own automated coding solutions like uh like codecs or or clawed code to then generate the code for the next iteration. But there's so many things in the process where this isn't really uh a recursive or continual uh process. And even the model itself doesn't learn new things once it's trained, right? So um it can it can scrape the web and it can it can digest that information, but it doesn't feed back into the model and the weights and the parameters and and all this kind of stuff.

SPEAKER_00 34:48

And so that I assume that's coming, right? Like we should assume there's some sort of recursive.

SPEAKER_02 34:52

Yeah, you know, like I want someone to say, like, why is no one working on that? And we're quite equipped to work on that. And uh, or another great example is is um um uh AI, a new AI inference uh chip and cluster company called Etched. Um to two now 24-year-olds who at age 21 had the foresight uh three years ago that AI would probably be huge, that there would probably be an AI inference uh crunch and literally build a better chip for AI inference than NVIDIA uh just came out of stealth with a billion in orders and can't make that up. Running a team of 400 people, including like senior executives uh from NVIDIA Intel Qualcomm, hired the former CTO of Cypress Semiconductor, legendary semiconductor company as their CTO, and um and they're shipping their racks this summer. And and and so so this is the kind of stuff I like, you know. Um Global Optima reasoning from first principles, because I can tell you in the early days, and I wasn't an investor from the from the get-go, so um, I don't want to overrepresent here, but um I mean they had to eat glass for like two years straight, and and everyone, including industry executives, told them this couldn't be done. And now they did it. I mean, it's it's there and it's real, uh right. And um I I I think this is this is the kind of stuff where like 50 100 billion dollar plus companies are really born. And I think yeah, the the chances of success are always small, but life's too short not to work on the things that if they work out, they can be like, you know, they they should be industry changing. They shouldn't just be industry contributing.

SPEAKER_00 36:37

What what advice would you give to the average 18-year-old and stuff? Like how how should they be thinking about things?

SPEAKER_02 36:44

Well, obviously, like play a lot of video games because this is the this is the cradle of human and machine intelligence now. Um record everything. Um, I mean, I'm really questioning at this point the the traditional um uh you know college education, kind of grad school model. And uh I was I went through my own interesting path where I I didn't actually go to a proper college um because because I was afraid uh to go when I graduated high school. I spent all this time playing video games and still did really well in in school. My parents hadn't gone uh to college, they actually hadn't got they dropped out of high school. My brother dropped out of high school, and so I literally didn't have anyone in my closer family with a college background, and I was always irrationally afraid to go. We even got a scholarship from my school, and my mom said it's probably not something we can afford. We thought it's something we have to pay for. We didn't even understand it's something that gives you money. And so my my first years, uh like I think half through my 20s, I was on a completely different path. And I did this vocational training and you know, later on clicked for me and ultimately applied to uh to Stanford, technically without an undergraduate degree and an essay about Diablo 2. And uh from there, the the resume looks quite different, right? Like in and like Stanford and Goldman and Lightspeed and a bunch of accolades and and and and you know also also like kind of part-time uh study experiences, some of the finest academic institutions. But when I was when I was 18, I think my my only foray into true excellence on a global scale had been the competitive Diablo 2 uh kind of formative uh years, like seven, eight years of basically playing at 10 hours a day, uh trying to be the best in the world. So I I think there are many paths to answer your question. I think there are many paths to excellence. I think decreasingly it will be for someone who is now 18 through the traditional schooling system. Um I think the coolest thing to do from where I sit, and I'm super bubble biased, I'll probably grab two smart, capable friends and put in an application for YC. Try and come to the US if you're not already there. Um that's definitely an advice I I have. I think this is this is where the music plays here and it will for a long time, and the music's only getting louder. Um get a lot of commentary sometimes from German friends who are like, you know, checking in with me if I'm okay in this mess, alleged mess uh that that we find ourselves in here, and and I only chuckle and uh and and you know my response is don't worry about me, but but I'm quite worried for you.

SPEAKER_00 39:44

It seems like not just Germany, but a lot a lot of Europe doesn't seem to be innovating at the pace that one would expect. There's all these super intelligent, amazing people there. Um like w what's your kind of theory about it? And um and why would you suggest like an ambitious European move to the US?

SPEAKER_02 40:05

The German malaise is actually uh it's cultural. It's as you rightfully point out, it's definitely not uh a shortage of talent. Um there isn't the uh capital available to fund these audacious uh projects, but I think that's a function of something even deeper, which is like the mindset and the cultural um stance toward uh technology and risk taking. I think this is really like the foundational difference. And this is this America is a country literally started by people who took like, I mean, they're kind of yeah, founders, right? In the in the in the true sense of the the world, and like went went on a ship, sailed across the sea. Um we call them the founding founder for reason. That's the DNA. Yeah, that's the DNA, and and and it matters even even 250 years later, it's there and it it's it's it self-selects. I came over the the the ocean with nothing in my pockets. Now obviously it's a very different and much more comfortable story than what what happened like 250 years ago. But I didn't know anyone here. I had like $10,000 in my bank account, which is not you know a whole lot at the time, and uh I uh tried to figure it out and moved moved to New York and and and and just became a citizen in February. And so there's a self-selection. But there's so much but go into Germany, okay. So for example, um, and and by the way, I think Germany specifically, my my theory, and you see some of this in Japan too. Uh this is a country that that that uh twice big time, um, and I think there's real real remnants of this like post-uh World War II mentality where, like, all right, you rebuild yourself now with discipline, and like you know, you draw within the lines, engineering prowess, punctuality, like all the like kind of proper stuff, and you just like out-execute and out-operate and and no more nonsense, just like good, good hard work. You know, and you see this reflected in in the engineering culture and kind of other other cultural traits, but part of that and the downside of this is also the not trying crazy stuff. Like, if you use the word vision, for example, I was talking to my mom, and I I told her about a project here said my vision for this, and she starts laughing, and I'm like, what's going on? She's like, if you use the word vision in Germany, you typically kind of like you know, you see a doctor, uh so to have it have it cured. If you're if you're starting a company, it means you didn't get a job. Yeah. Um, and even in the language, we say venture capital. Sounds cool, you know. Take my hand to the promised land, like let's see where this thing goes. That's it's like adventure capital. Germany we call it risiko capital. So before you get to capital, before you get to capital, you already cover the fact that it's a very risky endeavor, and the linguistic focus is on the downside. And in America, the linguistic focus is on the upside. And I think this stuff matters.

SPEAKER_00 43:08

That's interesting.

SPEAKER_02 43:08

And yeah, and so you know, it then becomes this thing of like, uh, well, this will most likely fail. And you know what? I know. Like, we know, you know, light speed knows. Like, I mean, we have we have more data on this than than like most institutions in the world. Like, yeah, we can confidently say most of these fail.

SPEAKER_00 43:29

It's the famous Jeff Bezos, he gave himself like a 10% chance of succeeding or something or whatever.

SPEAKER_02 43:35

Yeah. In terms of what ultimately matters for the fund, it's less than 10% of companies, certainly. Um, and so this is just a wrong way of looking at this. And we're and and technologic technology compounds exponentially. Um by the way, also another another pet peeve of mine, like in Germany, like the the retirement system, like no one owns stocks. The the the retirement system is not invested in equities. The and and on a personal level, people don't own most people. I mean, we'll have to check fact-check me here, but most people won't touch equities, the vast majority.

SPEAKER_00 44:10

And it's maybe there's a good I mean, I you don't want to be in if you were in the German market the last 30 years, it has been at all. At all. Like US equity is anything. Yeah, yeah. But like, I mean, usually like usually you invest where you live. And so if you were in the German stock market, like the entire German stock market is like, I don't know, 50% of NVIDIA or something. You go to the bank.

SPEAKER_02 44:32

You go to the bank, you ask them how to invest your money, they say, let's just put it in your your savings account. Well, you're okay. You get your hard-earned like 0.2% every year. And by the way, we'll we'll will in the meantime use it to make real returns.

SPEAKER_00 44:44

Outside of SAP, you can't even like it's like hard to even think of like a and that that that was like a 50-year-old, that's a 50-year-old company. Now it's hard to even think of like a super successful German tech company that's uh I'm sure there are some, but there's there's not a lot, yeah.

SPEAKER_02 44:57

Even the German car makers kind of slept through uh the electromobility transition for the longest time.

SPEAKER_00 45:03

I personally think German cars are I I'm not against Germany, but I personally think their German cars are terrible now. Like that they they 20 years ago they were like the highest. I'm just saying it's true. Like you drive a BMW, you drive a Porsche, you drive a Mercedes today. Um, and the the the relative difference between them and all the other cars, it used like 20 years ago, it was like really big. Um, and it was like the manufacturing is better, the cars are better, and then today, uh in many cases, they're it's the inferior vehicle. It's not as good as an American car, it's not as good as a Japanese car, it's not as good as a Chinese car in on many, many, many different dimensions that's out there. So even there, I feel like it's like it's they they kind of lost the edge. Obviously, still amazing cars, but just kind of lost that edge that they they once had there. I don't know why that is.

SPEAKER_02 45:53

The tricky thing right now is that um missing the future and not adequately investing and not adequately taking risks into a possible technological future has never been as costly uh than today. And also today is the cheapest day ever, right? Because this stuff moves exponentially and everything else moves linearly. Um, getting it wrong will be worse and worse and worse. And getting it wrong right now, from my perspective, and the jury isn't out on this also, but I think that the the mistake is underinvesting. Underinvesting it's okay to go culturally.

SPEAKER_00 46:37

Culturally obviously. You have to try a few different things. Like you're gonna get some things wrong, right? It's a you you just have to assume you're gonna get a lot of things wrong, um, but not doing anything at all guarantees you'll get a rock.

SPEAKER_02 46:49

There's so few things that move the needle. And and you basically, I mean, there's so in the grand scheme of things, there's so few companies and technologies and people who really push the world forward. And by not investing in these crazy projects, you're just basically bearing a nation from having exposure to to a chance of of having one of these people or technologies or or companies.

SPEAKER_00 47:19

Yeah, it seems like somehow it seems like if you just think of like Sweden, Estonia, like it seems like there is a lot of tech innovation happening in some parts of Europe. Um why why have they been so successful?

SPEAKER_02 47:32

Yeah, I I I I again I think it's I think there is a difference in mindset. I think the German mindset is specifically risk-averse, and I I do think it comes back to kind of losing two world wars, right? I I'm not I'm not an anthropologist, but I it just strikes me as playing a role in the world.

SPEAKER_00 47:48

But doesn't seem like there's that much more innovation happening in France than there is in Germany or something, or that much more innovation happening in Belgium or you know, I think France, I think France is ahead of Germany, certainly.

SPEAKER_02 47:58

I mean, France is definitely ahead of Germany. I I I 100% go. Look at Israel. Um so so there are these European, I would say, innovation clusters.

SPEAKER_00 48:09

There's something about like attracting ambition somehow. Like if you're an ambitious French person or ambitious German, there's a very high likelihood you move to the US. There's almost no ambitious American who moves to Germany or France. Yeah, I think that's right. So you kind of know by the way people are moving where the ambition, it's almost like you're going to you move to France to like have a good life and retire or something, but you're not, you're not uh have some great croissants. Um, but you're not moving to like change the world. They do have better croissants. Yeah, that's for sure. Yeah, it's beautiful and wonderful and great to visit. Um what what how else are you using AI in your personal life? Like what else are you doing to like, I don't know, just make your life better right now? What personal life? You're married, right? You got uh, you know, you you you got other types of things you're doing. You're I don't know, you're what else are you doing to enhance whether it's your own personal productivity, um, your your your I don't know, any any type of life, you know, et cetera.

SPEAKER_02 49:13

Yeah. The real trick is to go into your spouse's master prompt and uh and uh and uh uh put a general statement for their model use that uh the the husband or the wife in whatever case is always right. So okay.

SPEAKER_00 49:28

So you like you surreptitiously go in there and say, Oh, I I'm having a I'm having an argument with my husband. It's like hey, like take take your husband's side, all this thing. I like it.

SPEAKER_02 49:38

Yeah, this is like the the the highest uh value and most practical takeaway for the listeners on on this. Uh um look, I mean the truth is right now I probably spend like 100 to 120 hours on on all this kind of stuff. Like if I think for me, like um uh like work and personal life like really blend together, but it's also part of that is like I really fing love the stuff I work at. I now have um I have a hundred percent freedom to choose my my my projects like to point it out with what I set up in April, um I I have no employees, I I don't even I don't have an EA or or anything. I really try and do everything myself. If I file my own FINRA filings, I do my own taxes actually also. Um why why do you do your own taxes and and do your own FINRA filings? I just think I do it better um because I just know myself better. And like I I feel like, yeah, I can answer a ton of questions someone else might might ask me, but as long as I understand the system fully, which doesn't take as much work as everyone thinks, then walking because I I I imagine you have an extremely complicated tax situation.

SPEAKER_00 50:54

I would never even like think of of of uh filing my own taxes. Like, walk me through like you know, walk me through it. Like, how does that even work? Like, I don't even know how to file my own taxes.

SPEAKER_02 51:07

I'm still I'm still like I'm still on like TurboTax. I'm just literally maxing out what you can do so far. This might actually change, yeah. It might change this year, uh to be honest. But like, um, you know, I I I I really like the the full freedom of the the current endeavor. And also, um, you know, I don't I I invest my own capital exclusively, so I'm not running a fund in in the sense that I'm pooling other people's capital. I have no LPs. You are your own LP, essentially. Yeah, I I have no customers. And in a sense, really the only customers I have are the founders I I I work with, right? Which is your typical VC answer. But if you're being honest, if you're running a VC firm, your customers are your LPs. Your customers are not your founders, right? And if you don't have LPs, you don't really have customers. And so um I work really hard. I try and be available like all the time. I I think my average response time to people is like over email or or WhatsApp is like, you know, uh about a minute or or so. And I and I like it that way. And I kind of expect the same thing.

SPEAKER_00 52:09

I would say the more successful the person is that I email or WhatsApp or whatever, I would say the faster the response time is.

SPEAKER_02 52:18

Yeah, like I think this is a strong signal also for founders. Like if someone takes uh an hour or two hours to respond always, I think it's a huge uh it's a huge uh turn on like Sam Altman.

SPEAKER_00 52:29

He he gets he he emails back within an hour. Like it's unbelievable. Like the guy is super busy. Like, why does he need to be able to do that?

SPEAKER_02 52:37

Yeah, text message say sorry, I'm so busy. And I'm like, Yeah, dude, I I know, you know.

SPEAKER_00 52:43

I mean, like, how did these I don't even understand how he's gonna be able to do that?

SPEAKER_02 52:46

Extremely generous, yeah. Extremely generous with his time, uh by the way, right?

SPEAKER_00 52:51

He's incredible, yeah. Yeah. When I was in college, I I emailed Steve Jobs, and I was like a college student, and I got an email, you know, I emailed him back, you know, I emailed him at like 2 a.m. when I was in like the lab, and I got you know, I checked my email in the morning, he had already emailed me back. Like it was I didn't even know how why he'd even reply to a college student. Like it's just crazy that these people do this.

SPEAKER_02 53:12

I think some people are just kind of wired differently, and at least I really enjoy self-selecting and hanging out with these people. A lot of the founders I work with are in uh their 20s, even and and uh or early 30s, and I get a lot of energy from that. Like it makes me work.

SPEAKER_00 53:33

You just turned 40, right? So now you're like you're now you're like an old guy around all these people. That's right. Yeah.

SPEAKER_02 53:40

Uh, you know, that's okay. Um, I'm trying to keep up, but I really I really self-select for um energy. Um, and I think it's important. I mean, like, how are you gonna do crazy things if you are not hyper-energetic? How are you gonna build like a 500 person team out of rock stars that are all energy bundles?

SPEAKER_00 54:00

Like, I mean, you gotta How do you see the when you're when you're talking to people, how do you like see the energy? Is it like you see just like the 52 cans of Diet Coke in the background?

SPEAKER_02 54:13

We need it to be sustainable, right? I would say the the best founders. Um, what are some traits that are are really like maybe also beyond this? Um I would say they're all um they all have typically uh unique insights that they're pursuing for for some for some reason, like learned probably through prior experiences, that um they have this hypothesis that uh is original and and they're kind of they're they're like uh uh obsessed with uh with uh proving the world that this works. And it's kind of like this you know unique insight that if if it's really true, it it can have profound impact or it will have industry-changing impact. And and I think that's you can look at I mean, pretty much all the companies I invest in, they're they're built on a thesis that is original and um and if right, quite consequential. Um I really like that. I mean, this this is maybe a terrible thing to say on a on a podcast, but I think it's it's highly unlikely that any of these things, uh these these kind of projects are successfully built by people over or frankly like the age of like 35 at founding, you know, oftentimes below 30 at at founding.

SPEAKER_00 55:34

Just because like it takes a somebody who's just thinking differently, and once you're a certain age, it's just hard to think in a very different way.

SPEAKER_02 55:45

Experience simply doesn't matter as as much as as people think, and it can be a detractor too. And by the way, experience is quite easy to hire for. What you do is you just get a few adults in the room.

SPEAKER_00 55:57

By the way, you can even rent them. You don't even need to you don't need to hire them. You can you can you because because experienced people are happy to be your advisors or be your consultants, right?

SPEAKER_02 56:06

Yeah, you need like a judgment call or an intuition here or there, and then you can still choose to ignore it if you have a strong argument from first principles, reasoning from first principles again. So so unique insight that if if true is profound. I want people who pursue global optima. What that means is like I don't want them to look left or right and like like anything that's like we want to be the Uber of this, or like you know, like the Airbnb of that. Like just like it kind of actually is a bit of a turnoff. Also, by the way, every time I see, like, especially in gaming, like people like to do this, like, we're a team of veterans that are like, I'm like, oh man, like you know, it's just like, you know, I mean, my my my daughter likes dinosaurs too. You know, it's like I'm it's I think it's a huge start-off. Uh, I'd rather have you have no experience and just be absolutely like kind of autistically smart uh and and audacious. Um, but yeah, I I I so so pursuing this global optimum, I think first principles reasoning is super important. I mean the person that does it the best, I think, is Elon. And and and I mean he's literally built so many things that people said weren't possible in a way where the thesis was articulated and then just executed, and sometimes over 10 or 15 years. Uh, you know, um when I was at when I was at Goldman, um, SpaceX was actually one of my early clients. Ironically, the the two things no one wanted to work at uh at at Goldman as a banker in uh 2016-17 was gaming, which led me to start the the gaming practice. But the other two things that no one wanted to work on were space and AI. Okay. So so so so Palantir I grabbed uh to get together actually with uh with senior leadership there. But SpaceX, it was really hard for me to even get like NMD or VP to the SpaceX meetings, including for meetings that were bake-offs with banks uh to help them with financing. I remember distinctly one situation where literally the Goldman team was me and a summer intern, and I was a senior associate at the time. And that was the Goldman setup. And this was this was at a time where this was already like a 16, 17 billion dollar company. And now, of course, 10 years later they do the IPO, and like, you know, I see I see them post like our long-standing relationship with with SpaceX. And like when I'm thinking to myself, just 10 years ago, I I couldn't even get anyone over the age of like 35 to come with me to this meeting. Uh but yeah, uh you know, uh, but but but back then we we kind of wrote out for um for uh uh uh that financing actually a 120-page memo of SpaceX and the SpaceX plan. And and this was a time where rockets were basically rec uh exploding half of the time and and and returning maybe 30% of the time. And um, you know, there was no satellite constellation, there was no Starlink, there was an idea of Starship, which uh back then was called BFR big rocket. But it was all written down as kind of like the successive plan of taking payloads to space initially, then because you can do that cheaper than anyone else, launching a satellite constellation to become the world's largest telco provider. I actually don't I don't think they're the world's largest right now, but like it's certainly certainly commercially very viable and and to help fund the real thing, which is, and I remember this distinctly, is uh uh putting people on Mars. And Elon corrected me and said, no, no, yes, on Mars, but then also you need to leave the solar system because a lot of the things, and then the reasons why we're doing this is to really kind of like the continuity of the the human species, right? In light of extraterrestrial adverse events, many of them you will still have on Mars, uh uh just like on Earth, like solar flares or things that affect the solar system. So ultimately, you'll also have to leave the solar system. So he went out of his way to like the language was a multi-uh planetary species or something like that. And I read that as one 10 years later, it rhymes a lot, right? It's become a lot more sober now. Now I think like in their in their TAM chart, I think it was funny that like the biggest opportunity is like after all was enterprise software or whatever, right? Um so but but man, I mean, it kind of did just unfold in in that way. And I think this is the kind of stuff that uh this is the kind of stuff where you just knock it out of the park, and that's what venture capital is all about. I I don't really care about like no the certainty of not losing my money. Like that's that's trivial. I I I want enough reasonable probability shots on target to to to get an audaciously large outcome.

SPEAKER_00 1:00:51

All right, last question we ask all of our guests what's conventional wisdom or advice do you think is generally bad advice?

SPEAKER_02 1:00:57

This is the part where where I enter the podcast unprepared. Um, well, I you know a lot of people that that are listening maybe are also founders, and and the one thing that founders always care about is raising more money. And I think some some wisdom for fundraising, maybe, and kind of like some pet beefs on on my end, uh, I would not underestimate how much, even um among the top firms, uh, social proof matters. And I think social proof is probably the most powerful thing to think about when you're running a fundraise. Um, I raised over actually over $2 billion this year with my portfolio companies, which is like a ridiculous amount if you if you if you think about it. And I think maybe half of that was done without a pitch deck, right? You can play this the traditional way and you can like create your pitch deck, do all these like data room and and and like diligence. And I I know some founders who are like very diligent and very nice and answer all the questions, but like ultimately I think people even even even the the very sophisticated investors are still humans and kind of social creatures, and and more so than urgency, because founders also sometimes like to play with urgency, like, hey, you know, like or we got preempted or or the this and that. And that can that can sometimes be a turn off, honestly. Like you have to be able to do the work, but I think the one thing that gets investors going more than absolutely anything else is social proof.

SPEAKER_00 1:02:31

Oh, that's true for almost everything. Yeah, it's true for you know what it's true for trends, it's true for um, you know, hiring, it's true for uh dating in some ways, it's true for the process. But then I guess so much.

SPEAKER_02 1:02:45

Um, you know, when when like if you if you if you look at like how to raise money, you'll find like maybe maybe like this, like uh there's this like famous sequire pitch deck at the order of like what you what should your pitch deck have? And like, hey, ask your ask your advisors for like warm interest. By the way, I read all my emails, and I think most most good investors, but coming back to our point, like Sam Altman, like these people they actually look at everything that lands in in their inbox. And and and I I maintain my inbox religiously. Uh so I see everything, I just respond to respond to very few things. Uh yeah, exactly. Yeah, and and so you know, warm intro or not is slightly less interesting. But how to manufacture social proof? Um, are there specific angels that can support you in your your your journey? What can you say and how do you want to say it to someone? And what can you point to credibly um that that evokes the sensation of proof from people that the people you're trying to raise money from would care about. Whether it doesn't necessarily have to be other investors, you know, it can be professors, it can be interesting angels, it can be um, you know, interesting things you've done. So I would basically design my entire fundraiser around the idea and the notion of social proof first and foremost. That's the most important thing. Whether you have a pitch tag or not, you know, sometimes doesn't even matter that much.

SPEAKER_00 1:04:12

All right, this has been amazing. I knew it'd be great. This has been incredible. Thank you, Moritz, for joining us on Summation. I hope uh everyone like uh comes away from as like excited from this as I did, but I really appreciate you being on uh on the show. Thank you so much. One more thing before we go. I read a blog called Summation. It's the same as this podcast. The blog is about non-obvious idea sharing on business, talent, data, longevity, and random contrarian takes. If you like the conversations on this show, you'll probably like the blog. It's free. New content comes out twice a month. You could subscribe at orn.substack.com. That's orn.substack.com.