Moritz Baier-Lentz: How AI Is Rewriting the Rules of Gaming

0:05Hey, everyone, and welcome to Generative Now. I am Michael Magnano. I'm a partner at Lightspeed And this week, I'm excited to share a conversation that's all about gaming and artificial intelligence with my colleague Moritz Bayer-Lenz. Moritz is a partner here at Lightspeed, where he backs breakout companies that fuse AI with game design, storytelling, and world building. He was a world champion player of Diablo 2, turned Goldman Sachs banker and Stanford scholar, and now Moritz helps shape how we'll interact, create, and play in an AI-powered world. Here's my conversation with Moritz. Moritz, how are you?

0:44I'm great. How's it going? Good. Good to see you. You know, I see you almost every day. So just another great day here at Lightspeed. I mean, almost every day is probably an understatement. That's true. It's every day. It's probably multiple times every day. But, you know, we're always talking about deals and companies. We don't often get to talk about each other. uh and so i'm very excited to talk about you today it is one of my favorite topics it turns out for for people that know you um anyone that's listening that that might know you um or watching they would probably agree with that um i think there's no way to start a podcast with you without talking about how you got into gaming so let's do it so i think my gaming career uh technically goes back about 20 years now.

1:33And it's been on and off. But it started as a player, a professional player during my middle school and high school years. I grew up in a small town in Germany and my parents had dropped out of school and my brother had dropped out of school. And frankly, the bar was relatively low. They just asked me to finish school and I thought it was well doable while also playing a bunch of video games. As soon as we got connected to the internet, which was late, I think around 2001. Fell in love with Diablo 2, popular Blizzard game at the time. Played that every day. I mean, at least for eight to 10 hours, more on the weekend over years.

2:06And ended up peaking out at global number one ranking twice in 2003 and 2004. It was a wild west days of item trading as well. I mean, this become a topic in the Web3 era, where we're obviously also active with our blockchain infrastructure team. But the truth is that digital item trading was already a thing in the web one. Even before we had social networks, YouTube, Twitch or any of that, it was possible to sell off digital items on eBay, which was back then also in its first and second year in Germany, respectively. So, you know, very loose. You would find someone online who was willing to trade off their disposable income for your skill and time.

2:49And there was a market. And with that money that I made during those years, I ended up financing college and grad school, went into tech for a few years, and then picked up gaming many years later at Goldman when I was supposed to do tech investment banking, but had the idea to start a gaming practice for the firm. So hold on, take it a step back. You made money by basically going finding items for people like they would pay you to go find an item in Diablo 2, like a bounty hunter in Diablo? So basically, there were certain items in the game. And I mean, the big beauty, I think, of Diablo 2 was this beautiful systems design and items catalog design where, you know, as much time as you spend and put into this game, it never felt like you were done.

3:37You could technically finish the game in a day as in beating all the levels and beating all the bosses or whatever. But it felt like you could never finish this game and you would inevitably always make progress towards something bigger, better, improving your character, improving your gear, improving your skills and all of that. And so the ceiling was limitless, but I would say rarity and value of what gave you an incremental advantage grew exponentially the further you went in. And so it was incredibly hard to obtain certain items because there were, you know, it was very hard to get to the bosses that would even drop them.

4:18You would have to do it many, many times to get lucky on the exact right role of random attributes on such items. And so, you know, there was basically a translation of skill, knowledge, and frankly, time to find these items. And then you had some rich folks, you know, all the way from the US to Russia who didn't want to spend a couple thousand hours to get the best stuff in the game and maybe had, let's say, five grand laying around on the bank account. And they would take the leap of faith to basically wire the money up front on eBay through PayPal, agree on a time and day to meet in game, open a trade window.

5:03And for me to put that thing into the trade window and effectively traded for nothing because I've already received the money. So that was how it worked. Does the money get held in escrow or something like, or they're totally just trusting that you're not going to scam them? It was totally paid up front. And the kicker was also, because it was web one and not web three, the IP, including the items was technically publisher property. So what you were selling was the service required to obtain the item, not the item, which obviously No one cared about the service and the time. People cared about the item.

5:37Technically, what you could have done, and again, this is my 14, 15-year-old me doing this, you could have run away with the money and probably totally be fine. And the money was paid up front. And it was a total leap of faith. And obviously, that improved with Web3. But the fascination and the idea and what created the value of these things in the first place, which was people caring about a great game and who they are and what they can do in a great game, that hasn't changed. so how if you don't mind me asking like how much were you able to make collecting items for uh rich people in the u.s and russia more than uh more than a teenager should make and uh enough to uh help with uh financing college and business school in the u.s that's wild cheap that's crazy what what like when you are ranked number one at diablo 2 one of the most popular games in the world like what is that especially it's all online and as you said this is web 1.0 we're not all sitting on Twitter all day, you know, we're not all overexposed on the internet.

6:37Like what types of doors does that open up for you? What comes along with being number one of a video game in that era? There was not much to do with it at the time. There was no, I mean, there's barely any professional esports infrastructure. So the way to make money with tournaments, prize pools, or any of that was just starting. And I think, you know, that was really only the case for Counter-Strike, maybe with ESL, Electronic Sports League, forming at the time, other than item trading. And we saw the same thing happening in RuneScape around the same time. Like that was basically the way to monetize.

7:17Just to spell it out, there obviously also was no way to turn eyeballs, kind of creator monetization streams around this. No Twitch. Yeah, pre-Twitch, I think around the start of YouTube. But those channels didn't exist. Nobody was watching. No one was watching. Yeah. And I think even like, you know, async game watching on Twitch, on YouTube wasn't a thing yet. Yeah. Wild. It's crazy how different maybe things would have been for you if you were doing that today. I still like to play. Rocket League is my game of choice. you know maybe in the top one or two percent but you know top one percent is very different from top one there's a big big big gap in between you know right top one percent is like best among your town on on weekend pick up soccer and top one is is Lionel Messi or Cristiano Ronaldo and and right there's still some room in between but it's it's frustrating I'll say that like I could feel how things got harder even between age call it 18 and 22 i think our raw intellectual capabilities reflexes and all that kind of stuff peaks around age 20 and then it slowly degrades from there and we make up for it you think it's that and it's not just that more people came in most you'll find very few esports athletes over the age of 25 huh okay i didn't know if maybe it was just it got more competitive it got bigger it got more popular bigger pond that is also true and like you know the meta and and the frontier in these games have moved forward and people are now spending a lot more time and are willing to make it a career because it is more viable obviously you know everyone back then was primarily playing for fun like no one was really playing for money the money was a side effect got it that's wild so yeah i mean i knew i knew you were number one or you know highly ranked i didn't know the full story but you know when i met you i knew that and obviously had a sense that you were a super competitive person.

9:17And then I think within the first weeks of knowing you a couple of years ago, you're like, I'm going to run seven marathons and seven continents in seven days. I don't know, that seems impossible. So I feel like that's another thing. We just got to hear the story of. The main lesson for me from the video gaming days is. And a lot of this I realized only in hindsight, maybe 10 years later, that it was this beautiful global playground where so many opportunities presented themselves. There were so many ways to play this game. And to be number one, you basically had to deeply understand all the systems, take the entire thing apart and put it back together in ways that no one else was thinking about to kind of get that edge.

10:07I think because it was during my formative years, that's that ingrained something in me and this notion of arguing from first principles, chasing global optima, not getting too hung up with how the game is played, translated into looking at life in similar ways. And I think these are these are ideas that can translate into how you approach your job, maybe academia, but really taking a step back. I tremendously enjoy just laying out what are the next five years and experiences that I want to have. And it's a conscious exercise that I go through every December. And that includes obviously the things we work on here and how we set our strategies and focus areas, but it includes some experiences with my wife, friends, family and it includes things that i want to do physically how i want to educate myself how i want to grow spiritually and i think once you get a taste of the fact that the things you can do when you set your mind to them because i'm not a standout athlete i think my 23 and me has me marked for like tendency to overweight so i you know i i have to eat very disciplined and And I do a lot of sport, kind of still never see my six pack.

11:36Going into crazy adventures was one of the outlets for that. And it went from marathon to Ironman to ultras to this race that you just referenced, which is running seven marathons on seven continents in seven consecutive days. We started in Antarctica. Once we made the first step in Antarctica, we completed a full distance marathon there in South Africa, in Australia, in Dubai, Spain, Brazil, and then the United States, all within 168 hours. Two separate charter planes, a group of 50, and about two or three years of planning. And yeah, it is pretty crazy, but life's short. And I think this is my version of YOLO in the sense of, I think it's a precious gift and we should make the most out of it and life's too short to live it in a normal way.

12:32And this is, I think for me, this is one of the ways in which it manifests itself. How does it, Matt, how does that way of living and also your competitiveness and your drive now manifest into VC? You drew a connection to the marathons. Is there a connection to what you do now? Yeah. I mean, like, look, I think we're in a very competitive domain, right? There's, We're in the business of not only finding extraordinary people who build generational companies, but also convincing them that this is the right partnership and the right platform for them to engage with. And there's obviously formidable competition that we have.

13:13And there's a ton of smart people and a ton of motivated people out there who are trying to edge us out on what are the most enticing opportunities for upside. Not just for us as a firm, but I think at this point in the stage that we're in for the future of humanity in many ways. And it takes a lot of work and a lot of energy. I mean, you know this yourself, for us to stay up to date, to have original thoughts on the things that matter, which change quickly. I think to impress with the value that we can add in the right situations. And when we find something we like, you know, just like I do, what it takes to bring it home.

13:56And that is all consuming. Tell us about gaming, maybe kind of like pre-AI. Explain to the listeners and viewers like what VC looks like in gaming pre-AI. What is the opportunity? What are the types of bets you're placing or you would be placing? You know, what are the types of founders and teams VC is looking for in gaming? Gaming VC started in earnest, I would say around 2019, roughly in that time frame. Really recent. Relatively recent, yeah. before that if you had a good idea if you were making a game you would go to a big publisher get milestone based financing that way this is this is how kind of like video game projects more like indie game projects maybe got got funded for consoles for pc pc for console for the mobile phone um so when we see a new game come out on you know playstation 5 it probably went through this process yeah either it's part already of a big publisher so it is maybe a division of an electronic arts or take two or an activision blizzard or it is one of those independent efforts that receive publisher funding or milestone based funding um there are a few things that happened i think over the last i would say 10 years that made video game investing a lot more viable and palatable to the VC mindset, which is, you know, what you can't have is some big creative binary effort that if you get so lucky, makes a good amount of money in year one, everyone buys it off the shelf.

15:34Then you go back into year two, three, four, five of pretty much no revenue and having to come up with another game too. and like very lumpy revenue streams and kind of highly risky. You basically have to nail the launch. The most important elements here that turned it into something that looks a little bit more like SaaS software with recurring revenues is this idea of live ops or live operated games that are launched almost more like platforms. The game launches, but it's managed through the cloud in a way where you can push updates, where you can modify the game and effectively, you know, build this thing as you are in the market.

16:18I think Fortnite is a perfect example for that. If you think about, you know, how well Epic Games is managing it, in part also because they own the underlying engine, Unreal Engine, and they can just do a lot of magic that some developers don't even have access to. You know, you can bring an IP. You can literally, I mean, on some days, Fortnite looks like a car racing game. And then you suddenly tell your friends, hey, have you seen this? Like log in today. It's completely different. And, you know, there are special events, music events, film events, and even announcements get integrated into it.

16:49And it's this beautiful, I mean, people use the term metaverse, but it's this beautiful interactive social third place that can encapsulate all other media categories. Then as a VC, now you're looking at something where you have revenue in year one, but there's also a plan to grow that in year two and three instead of going back to the drawing board. And I think for us and the type of venture outcomes that we require, it's almost a fallacy to back games per se. I think the games need to look like platforms. Like platforms. Yeah. They're scaled consumer platforms with strong social network effects that even in year five plus still need to entice a player to come back, which in the attention game that we're in, not just within gaming, but competing with social media and all other outlets for attention.

17:42That is the war we're fighting. And then there's other things that help. Most purchases are now done digitally. They used to be purely physical, going to GameStop, which GameStop obviously developed its own complete dynamic. But digital distribution, live ops, these these are big big trends that made gaming a bit more vc investable and then obviously ai changed everything again so maybe still pre-ai um you know you're not necessarily looking for like the next elden ring which obviously like hyper celebrated game you know one of the best games of the year or whatever when it comes out but you're looking more for platform that can effectively like go on and definitely like fortnite where there's there are seasons like you know Every time there's a new season, it's like a whole new opportunity.

18:35And obviously, like you said, like a deeply social aspect to it. Yeah. So and obviously then, you know, I think for most VCs investing in gaming and interactive media, even before AI, the focus wasn't limited to studios, but it was also platforms and technology companies in and around gaming, things that can rhyme with Roblox, Unity, Discord, etc. You know, one thing that would be remiss not to mention is gaming is this beautiful playground that has actually brought upon a lot of technological innovation. I think the most perfect example is NVIDIA, which started as a GPU company for gamers. We built these GPUs originally because they were the most powerful engine for complex 3D calculations, which really mostly at the time was to make video games work.

19:25And then from there, it became effectively the most powerful computational backbone that we could come up with and brought us advances in autonomous driving. The blockchain all runs on GPUs. And then obviously, most recently, the AI and generative AI explosion. And now gaming is a fraction of it. And you could look at it with a similar eye at the engines like Unity and Unreal Engine, right? It's this very hard, complex environment, very competitive, very technically demanding. If you get something done in that space, it tends to have implications also, not only for other media and entertainment domains, but oftentimes in many sectors beyond that.

20:12So I think also with the lens of AI, it's not just that the gaming industry is using AI to improve itself. There are also precedents, even all the way back to decision trees, where gaming has also, through its own technological advances, advanced the field of AI. You know, again, going back to more of the traditional venture opportunities for gaming pre-AI, like what are these founders like relative to the founder that we know that builds, you know, SaaS product or, you know, consumer social network? I think everyone has an archetype in their mind for these types of founders. Is it similar for the gaming founders or are these people more designers, artists?

20:59Yeah, give us a sense. Leaving the studios aside for a second, if you look at the platform and tech companies in and around gaming and interactive media, I would say they look similar to what we would look at in other VC sectors. They are effectively building software companies or social networks of sort. And for the game studios, I think the key roles you'd be looking for are, I would say, designers and producers. Designers is not necessarily an artistic discipline. Designers as in designers of the fun, designers of the game mechanics, like what makes players come to this game. Like a product manager?

21:39Yeah, I think product manager is maybe a fair comparison for this. Like basically, these are experts in designing systems and designing engagement modes, for lack of a better word. And sometimes you even come up with completely different genres, right? There was the first time someone built a first person shooter or an RPG game or a MOBA. And so these are designers that come up with these decisions and test them out. And then a producer is probably more of a commercial manager, almost like a studio CEO, if you want. So I think the thesis by and large has been that in this field, you probably want to solve for people who have substantial experience, where maybe in other domains and in VC broadly, I think we're very open to hungry young entrepreneurs that have unique insight that if true have profound implications and are willing to try something new.

22:49AI has probably changed the playing field here a little bit as well and changed a little bit, you know, how people look at the industry today. But I would say that's probably a fair description of the last kind of like five years of early VC investing in the gaming industry. And in these teams, you know, so you clarify that a designer is not like it's not an artist. It's not a, you know, they're not designing pixels, but there are lots of those people in these organizations, right? I think art is typically one of the biggest proportionate representations in a studio and one that's going through big changes right now.

23:29Yeah, so that's a great segue. So yeah, now take us through the AI moment. Like, how is AI on the ground today impacting gaming, gaming studios and maybe even gaming venture investments when, yeah, I mean, I have to imagine these studios are now getting incredible leverage from AI, generative image models, video models, or maybe not, maybe not yet. Talk us through the transition that's happening right now. Yeah, I would say the early AI applications in the gaming industry were focused on, I think, productivity improvements, you know, including things like accelerating an ability to come up with concept art, even if you're not using stuff in production.

24:11you know the what's called the pre-production process of like finding your art finding your art style getting inspiration visual inspiration for engineers that is something that's probably most prevalent right now I would say there's still this is an artistic industry by and large and it's probably fair to say there's been a longer resistance inside gaming as there has been in other artistic domains to the benefits of generative AI. But what we can also clearly see, especially I would say over the last like 12, 18 months is that everyone's coming around to it. And there's a lot of pressure on budget reduction, especially for AAA games.

24:57And so I think it'd be a red flag for us if a studio wasn't excited about leveraging AI to improve internal workflows. And there are other domains here as well, like animation is one. where AI can help. But I think an important point to make here also, you know, as we speak about this, is using AI to improve existing player experiences is one thing. And I think it's a less interesting thing. And then the more interesting thing is actually to use AI to create what we like to call internally previously impossible experiences. In the end, players don't care about the sausage making. I mean, it's great if you shave some margin off by being more efficient, but players care about the sausage, right?

25:48If what comes out on the end is not differentiated and unique or blows your mind and you would say, wow, this is something we could not have done before AI. And some of those examples include more intelligent non-player characters, agentic simulations inside games, like creating these truly living, breathing worlds where not everything is pre-scripted. Objectively, it's not something we could have done before AI. That is actually changing the sausage, so to speak. And I think that's where the value lies and where we have to be adamant and focused because this is a learning we can draw from previous platform shifts.

26:23When the mobile phone came around, And the companies that didn't master it, and a lot of them were the traditional incumbents at the time, the big PC and console players, they all try to port the existing experiences to the mobile screen. After going on for probably way too long and saying, oh, no one's going to want to play on their mobile phone. It's too small. It's not performant enough. And once you get around to it, you try and do exactly what you did on PC and console and bring it to the mobile phone. at least two fallacies with that one is it's a completely different interface right the touch screen is fundamentally different from from from a non-touch keyboard mouse or controller controlled monitor or screen and also and maybe more importantly the ways in which players engage and with this new modality is fundamentally different it's much shorter session times you carry it in your pocket while you wait in line for the DMV, you pull it out and, you know, you complete a level and then you put it away and then you get a notification.

27:27You get back into it like a couple hours later for a few more minutes. That's so different from saying, I'm playing Diablo 2 now, sitting down and playing for four hours. Right. And so, yeah, we're seeing the same thing again. And I think we see enough now that I can say with confidence that I think history repeats itself here and companies and the startups that natively embrace AI for how they design their organization, but also the types of experiences that they're going after. Those are probably the ones that will build the multi-billion dollar outcomes. And it's our duty to focus on those.

28:05So it's not going to be the, you know, the classic studio that now just gets leverage on AI to make, you know, art creation more efficient. It's like, hey, what is the net new experience that leverages AI to do something that's never been done before. So you mentioned NPCs, non-player characters that are now alive. I imagine they're sort of agentic and sort of powered by these open-ended LLMs. What are some other new sort of formats of gaming that we're seeing or new elements of gaming that we're seeing based on AI or that are built on AI or AI first? I think the most interesting applications that we're looking at is one is anything that brings non-scripted behavior into existing game worlds.

28:51And we touched on that a little bit. I think that's one thing. And, you know, one portfolio company of ours that is pursuing the strategies in world AI. The other interesting pattern that we see with AI is enabling everyone to build video games. And that is currently unfolding in two fundamentally different ways. And one is early and the other is very nascent. The early one is Vibe Coding Games or a few companies that have basically developed prototypes for a prompt to game experience. And this makes sense if we think through the modalities that are already going through the prompt to medium transition right now.

29:33obviously text, images, and sound and music and video, we can now all generate. It would logically follow that the next frontier here to crack is to, you know, to enable users to also build generative games. And there are two different ways. The one where we see a little bit more progress right now, and it's also a little bit easier to wrap your head around, is effectively automating coding in the traditional ways in which game development happens today. So still in a game engine with textures and meshes and effectively automating the things that a designer or engineer does today. And that's akin to, let's call it like a cursor for game development.

30:24Just like we can automate software development, we can automate game development. And it's more complex and harder because it's not just code. That's one area that we see. The one that is very nascent right now and maybe even a little bit wilder is this idea of world models, generative interactive video, which takes the premise that we can now generate video. And ultimately, if you think about a video game and what we're looking at most of the time, at least until we have augmented reality glasses and things are truly 3D, what we get from a video game, what reaches us in 99 % of the cases is 2D.

31:09It's on a 2D screen. So really, if we could perfectly fool the player, so to speak, with a live interactive 2D video stream, it would be indistinguishable from what happens in the background of a traditionally developed game. Because you don't need to know all the 3D calculations that are going on as long as it looks coherent and what the final 2D output is. And so there's a few companies who are basically at the cusp of generating video in a way where it responds to your actions, key inputs, mouse inputs, controller inputs, and gives you the next frame that is logically congruent, including lighting and shading and multiplayer interactions and can render this quote unquote next frame at 30 frames per second, 60 frames per second.

32:07And right now it's still very low frame count, low resolution. You can also look at mid-journey pictures from like, what, four years ago where the dogs had like three eyes and seven legs. And I think our job is to squint a bit, but squint smartly, maybe, and then try and find analogies and or at least come up with the hypotheses that would need to be true for this to become a reality and in a reasonable time frame, right? Sometimes it's not enough for us to be right on something ever happening, like brain computer interfaces or augmented reality glasses. I mean, I think we can probably agree they will be part of the future.

32:50And the bigger question sometimes is whether it's in the next, call it two to three years or maybe 10. Just to get really clear, what you're saying is in these types of world models, if you could generate everything, the next frame based on keystrokes, etc., you know, all the teams and the organizations we spoke about, you know, the dozens or hundreds of artists like that just goes away because the model sort of does it automatically. Is that kind of what you're saying? I think the most extreme version of how this could manifest itself would be a future where you may no longer need a lot of the roles that make up a studio today or in an even more extreme case, studios altogether.