← All transcripts

How AI Could Bring Halo and Other Classic Games Back to Life Transcript, AI Summary & Key Points

a16z · 5 days ago · Science & Technology · 46:12 · EN

Watch on YouTube

Answer

AI could bring Halo and other classic games back to life through community decompilation and recompilation projects — four community efforts already made the 2001 Halo a playable WebGPU game — with platforms like Xbox now looking to support AI-assisted porting and remastering of defunct games.

AI Summary

AI is collapsing the time, cost and skill barrier between a game idea and a playable prototype. Astra can directly operate professional game tools — Blender, Unreal, Unity — letting a hobbyist go from game design document to concept art to rigged 3D models to exported playable game in one pass, and even QA itself from screenshots. Shorter dev cycles let studios take more creative risks and revive esoteric genres like RTS and job simulators that were never economically viable. Modding and remixing are becoming trivially easy — Counter-Strike came from Half-Life, League of Legends from Dota, PUBG and Fortnite from an Arma mod — and AI-driven decompilation is bringing dead games like Halo (2001) and Knights of the Old Republic back to life as playable web games. AI-native experiences can generate characters, abilities and quest lines on the fly, pointing toward the never-ending game: an MMO that adapts itself to whatever the player wants to do, in the way Ender's Game's mind game did.

Key Points

  • Astra's key unlock is direct tool use: instead of telling users what to do in Blender or Unreal, it operates the software itself, can QA its own game via screenshots, and enabled one-shot hobbyist game builds from design doc to export.
  • A second path to game generation is coding: a recent Claude Opus release was unexpectedly strong at 3JS, laying out in-browser 3D element by element in a way humans would find maddening.
  • Shorter dev cycles address the structural problem of 12-13 year production cycles like Grand Theft Auto 6's: cancelled games, suffocated innovation, and risk-averse greenlighting of sequels over new IP.
  • Lower costs cut both ways: cheaper games mean more creative risk and new IP, or the same dev time spent on far more ambitious scope — building a universe with GTA 6-level detail.
  • AI changes who can make games: games require coding, art, design and music, but AI can fill the skill gaps, producing an expected explosion in the number of game creators.
  • The last 20% of a game is 80% of the work — how a gun sounds, how the world feels — and that human taste and creative judgment won't be replaced by AI models.
  • Creation is consumption: most AI-generated games will be played only by their creator, and the zen-gardening pleasure of making a game is itself the point, as with Minecraft and Lego.
  • Anish built a Habbo Hotel-themed '90s Toronto record shop with Opus 5.5 that regenerated '90s genres and a working DJ; extending it to a whole street, his son added a pet adoption center and the agent rewrote the game engine to add a pet-following mechanic — dynamic abilities, not just dynamic content.

AI in practice

Used for

Agents

  • Savvy — An embodied in-game agent on the Spawn platform that accompanies players across games, takes bug reports, modifies the game live while you're playing, and creates portals into other games. 2 held 30:44

Business ideas

Use AI's collapsed idea-to-prototype time to build esoteric, low-audience games — like job simulators or modern RTS — that were never economically viable when a full dev team and years of production were required, and charge enthusiast players who will pay for exactly their niche.

For
Enthusiasts of a niche — people who want to roleplay a specific job or play a neglected genre — who will pay meaningfully for it.
Solves
Niche genres like job simulators or RTS have audiences too small to justify traditional AAA production costs, so the games were never made even though demand exists.
  • A Tokyo/JR subway train conductor simulator where you must stop perfectly or your paycheck is docked

Reinvent the dating-sim/otome genre so the love interests are live agents with encoded personalities who react in real time to your conversations and remember your day — replacing hand-written expansion characters with generated ones.

For
Players of dating sims and otome games, a genre currently generating hundreds of millions in revenue.
Solves
Content production is the bottleneck: each expansion pack is a new character needing a rich hand-built backstory, which is slow and expensive; agents could generate and embody them instead.
  • Love and Deepspace — a Chinese-made otome game generating hundreds of millions in revenue with hand-built character expansions
🔒  Build steps and tools for 2 ideas. Unlock

Tools & resources

8 items

BNo. 0935
AIAINotes.us Tool

Blender

Open source · blender/blender

Blender is a free and open-source 3D creation suite developed by the Blender Foundation. It provides tools for modeling, rigging, animation, simulation, rendering, compositing, motion tracking, video editing and game creation. Blender is distributed under the GNU General Public License.

Mentioned in
12 videos
Kind
Other
CNo. 0016
AIAINotes.us AI product

ChatGPT

chatgpt.com

ChatGPT is a general-purpose AI assistant that can generate movie ideas and expand them into story and storyboard prompts. It can also produce software projects from natural-language prompts, such as a Python dating application with user profiles and swiping logic, and help explain electronics and hardware-development concepts. It is also cited as a general-purpose AI capability that consumer products can package into more personal, character-based experiences.

Mentioned in
122 videos
Kind
AI
MNo. 7894
AIAINotes.us Tool

Minecraft

In the AINotes directory

Minecraft is a sandbox game used in the cited discussion as a host platform for mods and AI-built game experiences. The episode describes people rebuilding or remixing games inside Minecraft and considers how AI could generate game code, art, characters, abilities, and storylines for such experiences.

Mentioned in
1 video
Kind
Other
RNo. 7896
AIAINotes.us Tool

Roblox

In the AINotes directory

Roblox is a decentralized ecosystem of player-created games, discussed as a model for communities and experiences that persist through ongoing user participation. In the cited discussion, it serves as an example of how game creation and remixing can be opened to broader communities.

Mentioned in
1 video
Kind
Other
SNo. 7895
AIAINotes.us AI product

Spawn

In the AINotes directory

Spawn is a shared community and game-generation platform in which the embodied agent Savvy handles bug reports, modifies games during play, and creates portals between games.

Mentioned in
1 video
Kind
AI
TNo. 2097
AIAINotes.us Tool

Three.js

Open source · mrdoob/three.js

Three.js is an open-source JavaScript 3D library developed in the mrdoob/three.js project for assembling and rendering procedural scenes in browser applications, including WebGPU recreations such as playable versions of Halo, Fable 51 Worlds, Human Atlas, browser FPS worlds, CAD workspaces, and physics games. It creates scenes containing cameras, geometry, materials, meshes, and animations, with WebGL and WebGPU renderers and SVG and CSS3D renderers available as addons. Its official site and repository provide documentation, examples, tools, an editor, community resources, and a JavaScript module distribution.

Mentioned in
9 videos
Kind
Other
UNo. 7890
AIAINotes.us Tool

Unity

In the AINotes directory

Unity is a game engine used to build games, including a Japanese-style dating simulation with branching dialogue trees.

Mentioned in
1 video
Kind
Other
UNo. 7889
AIAINotes.us Tool

Unreal Engine

In the AINotes directory

Unreal Engine is a game engine used to assemble playable games. In the cited discussion, it is described as one of the game-development tools that AI agents can operate directly.

Mentioned in
1 video
Kind
Other

Links mentioned

🔒 Full analysis locked

Unlock more videos and the full analysis

A credit unlocks one video's full analysis for good — the build steps, the tools and how each was used, the methods behind every use case. Pro opens the whole library instead, and raises how many videos you can analyse a day.

Unlock full analysis — free

Transcript

Searchable transcript of How AI Could Bring Halo and Other Classic Games Back to Life — a16z (46:12). Search for a phrase, then click its timestamp to jump straight to that moment in the video.

Captions sourced from the original video on YouTube, published by a16z. The video, its captions and all related intellectual property remain the property of their respective owners; AINotes claims no ownership. Provided for research, accessibility and search — see the Transcript Notice and Copyright Policy.

00:00 Grand Theft Auto 6 is finally coming out next month. That game has been in production since like 2013 2014. It's like a 12 13 year development cycle. >> Crazy games have long been one of the most complex products to create. But because you can have a uh help build a role of some of these specialized knowledge, you no longer need to be a professional.

00:20 >> There is having an agent sort of look at something and try to recreate a game from scratch. There's also kind of the meshing and the merging of two games or two art styles. Multibillion dollar game franchises have been created using mods. Counter-strike, for example, that was a Halflife mod. League of Legends was inspired by Defense of the Ancients, Dota, which was itself a Warcraft 3 mod.

00:41 >> So, imagine what the next 12-year game will look like. >> AI is now making it possible to build sort of niche games like esoteric genres that previously would not have been economically viable. >> What's a game that each one of you has been thinking about making for years and now you finally can? Um, >> it seems like there's just been this incredible shift in in the gaming industry and also what we're seeing from from Zolo devs and just tinkerers.

01:07 Tim Sweeney from Epic has come out and said a bunch of really interesting things about AI. The CEO of Xbox, Asha, has also come out and uh sort of said that they're thinking a lot about the hardware and software interplay on on Xbox. And it seems like, you know, there seems to be a tide shift. So, yeah, >> want to talk about that. >> Yeah, maybe I'll start off and, you know, Anish would love to hear your your thoughts cuz you you've been building some games in your free time.

01:38 >> Yes, I have. Yes. >> Um, so I feel like the big there was there was a big technology unlock about a month ago with with Astra um because it was very very good at directly using um very popular game development software. And so Astro is very good at using Blender. it could actually like work directly within Unreal and Unity. Um, and in previous previous generations of models, they would just explain to the user like, oh, you should go and like do this in Blender or you should do this in Unreal.

02:04 But Astro actually could like operate the software for you. And that was a huge unlock. And so for the for the first time you could have a hobbyist um someone who's never actually built any games before um use Astro to draft the game design document, generate some concept art, move that concept art into Blender, have them build 3D models from it, like animate it, rig it, >> move it over to Unreal, and then basically uh add some music and sound files, which by the way can also be generated, and then you can export that

02:36 as a playable, and then you literally have a game that you'd one-shot using Astra. And then the other thing that's been very valuable is that Astra can also um QA itself and so it can take screen caps of the game and then it can learn from its own output. Oh, I need to fix this bug or this doesn't look right and say um that's why the initial prototypes that people are building on X are actually like surprisingly good.

02:59 And so I think that was a huge technology unlock. Um and I think that has a number of ramifications which we're starting to see in the games industry. And so the first the first effect is that it's just dramatically shortening the time between idea and prototype. And so that's something that Tim Sweeney talks about actually as um in some of his Twitter posts, which is that um AI is now making it possible to build sort of niche games or sort of like you know like esoteric genres that previously would not have been

03:31 economically viable. Um, but because he's shortening the time it takes to make a game, he can actually now just spit up a prototype of a a strange or weird game idea. Y >> um because it it costs less money, it takes less time. >> Um and I think that's actually huge for the games industry because I think one of the structural challenges in games is that it just takes a really long time to make these games, right?

03:53 It's like >> um I think you and I were talking about this earlier, but like Grand Theft Auto 6 is finally coming out next month and like that game has been in production since like 2013 2014 it's like a 12 13 year development cycle. >> Crazy. >> And I'm sure it's going to be an amazing game, but like by any other software development standard, right?

04:15 Like taking 10 plus years to make um you know HR software or CRM or it's it's kind of it's kind of insane, right? M >> and so um shortening the time from idea to prototype like let's see if we can get a 5 to 10 year dev cycle to I don't know like 2 to 5 years you know it's still going to take some time but I think it'll just enable the game developers to take more risk um have more at bats like take more experiments and and and just generally just increase the richness and diversity of games that are created which I

04:44 think is really exciting. >> Yeah. Yeah. I mean to build on some of those comments you know it's interesting because we've seen two sort of approaches emerge. So I actually think the first thing that really was applicable to gaming was the emergence of 3JS manipulation and I think it was Opus 47 maybe it was Opus 48 but it just had this property actually which it was not designed to have in it at least anthropic said it did not um which is it was awesome with 3JS >> which is in browser 3D and it's interesting because

05:11 the way that it does it sort of it's laying things out almost on an element byelement basis like in a way that would be maddeningly manic meticulous for a human. Um, so coding is sort of one path to game generation. And it felt like what Astra did really well to your point, John, was tool use, >> you know, could use Blender and Blender in the cloud really, really effectively.

05:32 So these are two like mega trends within models, one kind of emerging and one programmed that are converging to drive a lot of the um the acceleration in sort of game development. >> You know, the the thing it's funny you say like 12 years to make HR software. Like I think just the amount of time it takes is why there's so many sad stories in gaming.

05:52 >> Like think of all the game stories that are left untold, right? How many games were cancelled 3 years into a six-year cycle or 5 years into a sixear cycle, you know? >> The market has moved on, player tastes have changed. Like >> totally you go through these financial cycles. You go through these creative cycles, you know, where the ideas that were creatively interesting 3 years ago now feel flat.

06:12 So, it's just so hard to work on something for 13 years and see it all the way through and then have it be great. It's a miracle that it works at all. >> Yeah. No, for sure. And the other aspect of um these long development cycles um which I feel like people don't talk about, but I I think has a very real impact in the industry is that the long dev cycles like have a suffocating impact on just innovation.

06:33 Because if you're a producer at aaa game studio and you know it's going to take 10 years to ship a game and hundreds of millions of dollars, you're going to be very conservative in what you greenlight, right? >> Interesting. Right. >> It's like you're going to you're going to greenlight sequels or the new IP. I think we're on like the 20th Call of Duty installment that's Grand Theft Auto 6, right?

06:54 >> Yes. >> It's not a new IP. >> Um you're going to skew toward genres that have established demand, right? which is why we see so many first person shooters, third person shooters over some experimental genre. >> When you lower the bar for sort of um shipping a game, right, it's going to take less time. It's going to take 5 years instead of 10. Instead of hundreds of millions of dollars, it's going to take, you know, tens of millions.

07:15 It's it's still going to be expensive. >> Yeah. >> I think it just means that people can take more creative risk. Totally. >> Like you can actually ship a new IP. you can take a bet on some some game developers that have never actually shipped anything before and you can greenlight them and get them a budget to make something. And I I think that's I think that's a really exciting story for for the game industry and and actually is going to unlock a lot of structural innovation, >> right?

07:38 And then the other actually way to explore this new capability though is to increase ambition. So imagine what the next 12year-old 12-year-old game will look like GTA 7. You know, you were talking today about how intricate the city is in GTA 6. >> Yeah. Yeah. They have custom radio stations that they designed because they pro they were just like, "Well, we have a year and, you know, this team isn't doing anything.

08:00 Let's just record, you know, a bunch of radio stations and stuff like that." But >> it's incredible, right? But you imagine if they then spent 10 years with the current tools, forget about a city or even a world. You could build a universe with that level of detail. Why not? >> Yeah. >> Yeah. Yeah. Yeah. No, I I think that's absolutely right. It's like um uh just like expand the scope of games, the ambition of games, take more creative risk.

08:23 I think um I think AI can can literally create an abundance error in games, you know, that that wasn't previously possible cuz we can make games that have never been seen before. You can take more creative risk. Um and and then another another element that we haven't talked about that I think is equally important is that AI is also changing who can make games >> because um games have long been one of the most complex products to create right it's a true multi-disiplinary craft like you need to know how to code you need

08:52 to know art you need to know game design you need to know like music and sound >> um and so the >> execution for for making a great game is just very high like much more so and just snapping a great photo or like making a video. Um, which are hard, but I think it's just an just a level of difference there. Um, but because you can have a uh help filled the role of some of these specialized knowledge which you may not have, >> I think you no longer need to be a professional like you no longer need to have a full team of

09:22 people that are all like, you know, specialized in different roles and instead you can have AI kind of plug some of the gaps in your skills and knowledge. And I think that's democratizing game creation in a way that wasn't previously possible. Um that you can actually have someone with just an idea and some reference art work with Astra to sort of like iterate their way to a game prototype.

09:44 >> Like of course it's it's a prototype. It's not a fully playable game. Um and then many people won't actually finish it. Like they'll they'll just make the first level of the scene and then they'll drop out. But I think like some percentage of them will make it through um and then decide that they love game development. They actually want to turn this into a a job itself.

10:01 And I think we'll see an explosion in a number of game creators. And and I think that's really exciting. >> You guys must talk to founders who are either creating tools for these kind of wouldbe AI enabled game developers or they're trying to like work on AI developed games themselves. And like the thing that would seem the hardest thing to me would be just game balance and making something that's like actually fun and engaging.

10:27 There seems to be this kind of it's it's an art and it's a real it's it's something that does I think require human intuition and human knowhow and um kind of just like a a different kind of creativity. Uh what what are you seeing there in terms of what AI can do? >> So I think you hit it spot on. I think um and I think the AI model is not going to replace human taste and creative judgment.

10:54 I think that is still really really critical. Um and then there's a there's a saying in the games industry that the last 20% of the game is 80% of the work. Um and that's where all the magic comes together make something fun because the first >> the first 80% you could say it's just putting together a game that works. It's like oh okay great like >> like you can start the game you can there's a character in it.

11:17 It can shoot things. There's like monsters to shoot. Like I think that's the first 80%. The last 20% is so critical. It's like >> how does the world feel, right? Going back to your point in like a custom radio stations in Grand Theft Auto 6, like >> when you fire a gun, like what is the sound of the gun? Like the shotgun should sound different from the pistol.

11:35 And then when you use a laser gun, there should be a recoil for that and so on. Like all the little details are what makes a magical experience. And I think that is going to be the province of humans for the foreseeable future, right? It's like how does something feel? What do players want? Like how should this gun feel? Um and so uh I I think that's still going to be critical for for human taste and judgment in in those situations.

11:59 >> Yeah. Yeah. and and like like what is going to make me want to spend a thousand hours playing, you know, Grand Theft Auto or playing, you know, Balders Skate 3, these like incredible games that like like I I think it really is just this level a of detail, but also like I think a sense of like consequence and depth to a world and maybe maybe that is something that LLMs can actually like >> help extend.

12:26 >> Yeah. Yeah. Exactly. >> What do you think? Yeah. >> Well, so like I I agree. I don't know that it makes it trivial to create a really fun game because I think that's really hard and it's an intersection of many things. Some designed, some serendipitous, some even cultural, some social. So there's many things that go into making a fun game. I do think though in this new era sometimes creation is consumption.

12:47 >> Like the vast majority of games are ever going to be played by one person that is the creator in the future. And that's okay actually because I think that there's you know programming or creating games for the output for the artifact and then there's doing those things as an activity >> and it's actually very tranquil to make a game and to work on a game and create especially because it has the art element >> seeing the assets come to life you know seeing the characters say things you didn't program them to say like

13:13 I I really do think there's a sort of zen gardening aspect to game creation which will become increasingly popular. I think that's I think that's such a great point which is we're seeing that um you know creation is play sometimes right >> like we've seen that for Minecraft which is one of the the most successful and longest running games of all time like we've seen it even prior to video games and things like Legos where people the kids are just creating things for fun and sharing it with each other >> but I think

13:40 that is definitely the case with with AI today as well which is just the act of creating something bringing something to life is itself fun and then AI is just turbocharging what people can create and so you don't need to know everything about code and art and 3D >> that to make something that you can be proud of. Yes. >> And and have fun making >> 100%.

14:01 >> Anish, you you built this Hobo Hotel themed kind of record shop. >> What was what was that process like? >> Oh, well, it actually started with a conversation with a friend who's a DJ and he was saying, "Oh, all the online record stores look the same. It's just a list of songs." And we were talking about how when we used to go record shopping in Toronto in the '9s, going to the record shop was like going to the barber shop actually because you go there, it's filled with people.

14:25 You can't not have five conversations. There's someone at the back who's playing tracks really loud. There's mixtapz for sale. There's raves being promoted. It just the whole thing was like it was a real social environment. Yeah. >> So he said, "Hey, why can't shopping, record shopping online be like that?" So I thought, "Hey, that's cool. Let me try to actually create that."

14:44 Yeah. >> So, this was um with Opus 55. I gave it a shot. I gave it just a couple Habbo Hotel reference art components and it just >> it super exceeded my expectations cuz not only didn't it generate the room, it also somehow grabbed all kinds of '90s music. >> So, it had a creative R&B, a creative dance hall, a creative hip-hop, a creative like soul.

15:05 You could pick the track, you could look through the tracks, put them on, then the DJ at the back would grab the record, put it on. Mhm. >> People were dancing on a little. It was very intricate. So it was a total inspiration. >> And then from there I actually said, "Hey, what if I extended this and turned this into Queen Street in Toronto in the 1990s?"

15:24 >> Okay. >> So I turned it into a whole street. And the last touch, which actually I think was the the most sort of science fiction thing that would never have been possible, was I made it so that every one of the other storefronts were something that another player could come and build out just through prompting. But the interesting thing is that they could not just customize the art, they could customize the capabilities.

15:43 So my son went in there and created this pet adoption center. Um, except you could actually adopt the pets and they'll follow you around. >> Now, of course, I didn't design this like animal following capability in, but the coding agent updated my game engine to create that mechanic. So now, you know, in the past, it was really inspirational to have procedural level generation, >> but that's just dynamic content.

16:04 Now we're talking about like dynamic abilities. whole different mechanics, right? >> Yes. >> And did um uh has your son played uh Adopt Me from Roblox? Cuz that's >> No, something like that. >> Is it similar? >> It's one of the most popular Roblox games. >> Oh, maybe he has and that's how he came up with it. I'm like, this is so creative. >> For sure.

16:24 For sure. I think that's a great example. So So my son has also recently got into game development using using AI and um it actually started because he wanted to play a game. It's called uh Ghost of Sushima. Mhm. >> It's a fantastic game, but you basically play a samurai in Beetle Japan. As you can imagine, it's like very bloody and kind of violent.

16:45 >> And so he told him, "Actually, no, you're eight. You're not playing this like bloody samurai game." >> But he really wanted to play the game. And so he started talking to, you know, the chat GPT and asking him about it and then it's and then and then somehow he discovered that he he could build a Minecraft mod of Ghost of Sushima. >> No way. And then he used GPT to basically deconstruct the game because there were all these like wikis online to break down how the mechanics work, how the boss battles work.

17:11 Um, and then so uh yeah, so essentially CEX helped him build a Minecraft mod which he installed and then it's Ghost of Sushima except as a Minecraft mod >> and it works. >> It works. It's super well. He's been sharing it with his friends and >> wow >> now at this point he doesn't ask us to >> buy new games anymore. He's just like I I want to play this game.

17:32 I'm going to make it as a Minecraft mod. I think that's >> a lot of inspiration. Well, what I also love about that is that it says, you know, it illustrates the fact that as we think about how games get better, realism is not the only or even necessarily the most interesting axis for improvement, right? Just the fact that you can create a concept from two existing concepts.

17:50 We were talking about this that never would have existed. That's right. >> Is extraordinary. >> That's right. I think this is actually a good a good transition to talk about decompilation and and that just really feels like a a new trend. And it seems like they're kind of they're different strains of decompilation. There is kind of >> having having an agent >> sort of look at something and and try to recreate a game from scratch.

18:14 There's also kind of the meshing and the merging of two games or two art styles. Um, and yeah, I just I would love to kind of disentangle all of these different things that are going on right now. >> Yeah. So, I hope that there's two trends that are happening that are getting turbocharged by AI. So, the first is that, you know, the games industry has long been driven by modding and and modding is the concept of like you take a game that you already love and then you build like an add-on extension to it that changes the

18:42 gameplay somehow. Um, and then it becomes its own game. It's it becomes its own genre. And historically, some of the biggest games in the in the games industry have come from mods. And so, um, Counterstrike, for example, which a really big esports, that was a Halflife mod. Um, League of Legends, um, one of the most successful MOBAs of all time, uh, was inspired by Defense of the Ancients, Dota, which was itself a, um, a Warcraft 3 mod.

19:08 And then um my favorite one like Fortnite and and PUBG um were itself inspired by um a Player Unknown Battle Royale mod of Arma. And so multi-billion dollar game franchises have been created using mods. And so this is just this is the history of somehow some of the the best and and biggest games in the world have been created. And AI is making it really really easy to mod games right now because exactly like the way my son did.

19:39 Um, an ordinary player, someone who loves an IP or a game can basically point Astra to it and say, "Hey, like make me ghost of Sushima except add this, right? Like add box crafting. Add the relationship system from Animal Crossing." And then you can actually just make it and then see if it's fun. And like if it is fun then other people start playing it and it becomes like a genre in itself and and it kind of blows up in a big way.

20:05 So um and you know and some part of this is um going back to like the creation loop. It's hard for someone to create something new using a blank canvas. >> Yeah. >> So if you just put a text like a text prompt in front of someone and say hey like make me a game like you wouldn't know where to start right. And so that's why working within the framework of an existing game that someone else has made and then just say, "Hey, like add this thing, remix it with this other thing."

20:30 I I think it becomes very powerful cuz it gives you a place to start and a place to remix from. >> Yeah. >> Yeah. >> Yeah. It's it's interesting. So you mentioned modding. I think the two other areas that I've seen, you know, one are sort of decompiling such that you have the source code and the assets. >> Yep. >> And I think this gets really interesting for beloved but old games.

20:46 You know, you take games like Knights of the Old Republic for example, which of course Boware is out of business. And you know, Boware is never going to make another Knights of the Old Republic game. And you think about extending it or forking it. Like that's really cool. And then the the third maybe most futuristic one I've seen. And >> there's this guy Blendy who's been doing a lot of these amazing things on X BL NDI.

21:09 um where he'll take the assets from two existing games and then combine them into a third game, >> you know, which is just a totally outer space concept, but it works. And some of the games are really fun. Yeah. I was going to say on the um on the on the uh the retro games front, um I think what sparked the uh the Xbox Labs announcement when Asha talked about um wanting to support you know AI and sort of like new new frontiers of of development specifically that um there was just so many community projects popping up

21:39 that were taking Halo like the original I think it was like the 2001 like Xbox game >> classic >> and then they were decompiling and recompiling it for like modern like modern sort uh platforms. And so, um I think there were four community projects that I heard of that made Halo a playable web GPU game and to just basically play Halo in in the web now, right?

22:02 And um it's it's really incredible that and it's going back to what you were saying earlier about 3JS, right? Like I think 3D on on on the web is never been like more powerful or more accessible. And so, um, I think Xbox has likely seen things like this happen and they're saying, "Hey, like we used to have whole teams that would just take a Windows game and, you know, port it over to the iOS ecosystem."

22:28 >> Um, or take old games that are defunct and then sort of like try to remaster it. But now AI can help a lot with those efforts and you're basically taking an established IP and you're exposing it to a new audience like kids today that have never that have grown up and they've never actually had a chance to play Nights of the Old Republic. Yeah. But now they can.

22:45 >> That's right. That's right. And potentially on a console or on a PC or a device that's relevant to them. >> Yep. >> Yeah. I think it's really really cool. I also think that there's genres of games that have really high production cost or have historically had, you know, where you simply don't get the ROI. Like, you know, people love RTS games. >> Everybody has been obsessed with RTS.

23:05 There's been so many efforts to make a great iPad RTS. And like, we've never really really gotten one. And, you know, now I think we're going to see this complete renaissance of genres like RTS. >> Such a It's what I'm Yeah, I think Tim Sweeney talked about this as well. That's what I'm personally really excited about. I I love this really niche genre of games called job simulators.

23:28 And uh it's such a strange concept. It's like you basically play a day in the life of a particular job. It's it's typically not a glamorous job like you're not like the president or something. Um the one that I the one that I played most recently is um uh this train conductor simulator. >> Oh my gosh. Amazing. where you can basically be a conductor of um subways and trains in various cities around the world.

23:52 And so, um I've been a like a JR, like a Tokyo subway conductor. It's exactly as it sounds. It's not glamorous at all. You show up for work, you have a timet to make. You got to drive this train and you got to stop at every stop along the way and pick up passengers. And you got to do it perfectly where if you're like late on like, you know, stopping like the last like the last car doesn't make it in so not everyone can make it on the train and you get your your paycheck is docked >> like for that.

24:19 And so it's like such a it's a niche genre, right? Like I can't imagine that many people want to be train conductors. >> Yeah. >> But if they are now like >> you can make games like this, right? And people like me who really love it >> Yeah. We'll pay we'll pay a lot of money to to play these games, right? And so >> is it fun? >> I I find it really fun.

24:40 >> You know, it's funny if you if you look at like old computing magazines and you look at like the rankings of of top software, number one is usually like Visical and number two is like flight like Microsoft Flight Simulator like early flight sim games and like like I actually think like simulator as a genre just really has a hold on people. People People love Sims.

25:03 >> That's right. >> It's Yeah, it's just an evergreen kind of kind of game. >> Maybe there's like a particular type of job that resonates for various people, but maybe you never had a chance to discover it because the simulator genre just wasn't that big. >> That's right. >> But now you can act you can you can roleplay a day in the life of almost any job in the world that you that you can envision, right?

25:25 Like wouldn't that be really cool? feces. >> Did you guys ever play Untitled Goose Game? >> Oh, I I I know about this game. I know about this game. Yeah. Yeah. Yeah. >> It's a goose simulation. >> That's right. You pretend to be a goose and you go around. >> So, maybe I am into the sim genre, you know. Yeah. Mischievous goose. >> It was It was really big during the >> That's right.

25:46 Goose. I don't know what that says about me. >> Chaos. Actually, going going back and kind of like merging the conversation about RTS games, like one thing that I've seen are kind of like agent orchestration schemes that look like classic RT like there's the Age of Empire skin over your like agent orchestration. >> Yes. Yes. There's a Minecraft one too.

26:10 All the different structures or different quad code. >> Exactly. Exactly. What do you What do you guys think of those? Is that Is that like a a toy kind of rendering? Is that kind of just an extension of the fact that you know it does kind of feel like a game to be managing your agents like that or is that like like is that the start of a new sort of genre of simulator?

26:34 I mean, I think it could be, you know, like being the general manager of a line of business at any company at Credit Karma or Google or wherever else is it is a sort of sim, you know, >> you're sort of working with all of these teams who and you have a limited amount of control and you have incomplete imperfect information and then you have market forces.

26:53 So, >> I don't know that it's that different. I actually think that now you can trivially create these really ambitious interfaces. Yeah. So in that world you might do that and it might actually be optimal whereas in the old world the cost was so high it just was insane. It made no sense, you know. Yeah, that's right. >> Sure. >> I think it's um I think one thing that the the gooies, the graphical user interfaces that are inspired by games being used for agent management makes me think of is that um the average

27:24 consumer finds agent orchestration to be non-intuitive, right? like work working from a command line trying to manage your agents you know these using the primarily textbased interface that is non-intuitive and so I I do agree that there will probably be a more mass market visual interface >> that an everyday consumer can use to manage like >> a collection of agents >> and so um and whether it looks like Starcraft or Agent Empires a factorial that I'm less sure of the exact implementation >> but it does feel like we're

27:55 early in sort of like figuring out user interface for primarily a gent work players. >> Yeah. >> Yeah. So, John, one thing I'm really curious to hear from you on, obviously Tim and Asha have come out, you know, really supporting the kind of community around AI developed games, which is awesome. But do you think that their companies and ecosystems >> are set up to really capture the value from this trend or do you think that they're currently in an awkward position today knowing that they need to change?

28:24 >> Yeah, I think it's a good question. But I think um I mean if you just take modding as an example, historically the original owners of IP >> Mhm. >> have not benefited that much from the mods that have been created on top of their IP, right? And so I think the original creators of of Dota didn't really benefit that much from League of Legends when that was created.

28:43 I'm not sure if the original HalfLife developers earn that much from from Counterstrike. And so I think there is a gap in terms of business models that need to be figured out. And sort of if you have if you're really embracing AI as like a remixing, modding, decompilation, recompilation engine for that empower your player base with like how do you as a company benefit from that sort of commercially, right?

29:08 And >> um I think I think folks are still early in that. Like I've heard, you know, I've heard talk of folks like trying to figure out if there's like a a revenue share or royalty that can go back to the company and >> um therefore you're powering like creators. But, you know, I I think there's there's there there's a lot to be figured out there still.

29:27 >> Um I mean, the strongest form in my mind is Minecraft. you know, like Minecraft really benefits from the mods ecosystem, >> but it's interesting cuz we've interchangeably used the works rem the words remix and extend, >> but I actually think that, you know, all of the additions that extend the core game and maybe share a core foundation of community >> plus a base game that actually really benefits the IP owner.

29:50 Y >> whereas when you're forking the game, there is no benefit. In fact, you might even be splitting the community. >> That's right. That's right. >> Yeah. And I think it goes back to something which um I know you and I have talked about um separately which is um how do you create new forms of distribution >> for the next generation of AI AI games or games that are created using AI right and >> I think today like many of these games are still going through traditional distribution channels right they're >> they're

30:18 being posted on Steam on the app store and you know they're you know if you're making something in Roblox it's still within the Roblox ecosystem and I think there is an opportunity for a company out to also create a distribution platform. >> Yeah. >> Which natively harnesses and empowers all these AI game creators that are making stuff and then um and it doesn't really exist today because of the fact that all the games are going through traditional channels, right?

30:43 >> Well, you should check out Spawn actually. You know, Spawn, we're not investors, but I'm a huge fan of what they're doing. It's Jacob and team, and they've created this sort of shared community plus game generation platform. Um, and what you do is you work with this embodied agent named Savvy. So, she's in all the games with you. She can like not only take bug reports in the platform, but she can modify the the game while you're in it, >> like a co-pilot.

31:07 >> Yeah, that's right. And then you um and she can create portals for you, so you can kind of step into other games. Yep. >> And you can bring other players with you. Every game is multiplayer by default with the same community. Mhm. >> So, it's it's early, but it's the most impressive, ambitious effort I've seen at doing something like this, and I'm sure there'll be more.

31:25 >> Yeah. No, I think that's um I think that's exciting. And yeah, I'm a fan of Jacob and I've been following the the company for a while. I think um I think the broader opportunity for hey, like how do you create a community that basically persists between games, right? And so you're not just >> here's a couple of people to play this game and a couple of people to play this game and they're not interconnected.

31:46 I I think that is a a huge opportunity, right? And I think Minecraft and Roblox are great examples of this where once you're inside the Roblox ecosystem, if you ever get bored of one game, like you've beaten a game where there's like hundreds of thousands, if not millions more sort of experiences that you can participate in, right? And >> I think that's Can you contrast the two, John?

32:06 Cuz here's one thing I've noticed like Roblox all the games are extremely different. So it feels like the kind of core shared game, if you even call it that, is actually pretty thin. >> Mhm. Whereas in Minecraft, the core shared game is very rich. Do you think you need to have a really rich game or is it because it feels like the pure sort of portal based worlds without a shared core gaming mechanic are harder to get off the ground or is that not true?

32:32 >> I think having a I think the benefit of a of a Minecraft model where it's more centralized, right? like everyone starts off playing the same game >> and to your point you're forking you're modding it you're making you're making it your own when you um create create your Minecraft mods I think the benefit of that is that everyone has a shared sort of experience and they're familiar with the IP and you're kind of departing from that and so like the Minecraft mods are almost like fanfiction right like >> but everyone's

33:01 like watch the same Game of Thrones series right and so they know who Daenerys and Jon Snow and what the watch is and you know and so on so forth and Minecraft has a lot of shared culture, right? after the enderman and like the dragons and you know um and that feels very different from a more decentralized approach which is what I would characterize Roblox as like cuz people can enter the Roblox ecosystem from any game like I mentioned earlier like Jailbreak there's there's a bunch of other games like Brook Haven >> um

33:31 and then you have a very different experience and then you have to learn like where to go from there and how to be a creator but both approaches are are totally viable right But um I do think um yeah I I I do think it's just a philosophical difference in terms of like how you want to extend uh the game and harness the community >> and both can work.

33:51 >> Yeah, both can work for sure. >> Uh going back to our discussion about studios and publishers, what do you what do you think they're the most excited by and what do you think they're the most worried by right now? I mean, I think getting their costs down, >> having like has got to be super exciting. Yeah. >> And just being able to create, you know, more IP with less financial risk.

34:11 Yeah. >> Seems to be hugely exciting. Look, for a lot of these studios, you know, I think even the studios, they kind of live and die by winning the right to develop a game for a publisher and then if the game doesn't do well, they may not get it again. You have hundreds of people that are overworked like crazy. It's got to be one of the worst qualities of life >> just being in that industry.

34:32 I mean, people do it anyway because they love it. >> Um, but it's tough, you know. So, I I think that just the quality of life of the average person in the industry and then honestly the quality of the business should go way way up. >> Yep. I think maybe what is worrisome to them is I'm guessing that you know Xbox I guess Microsoft owns Minecraft but they wish they had an asset like that that was closer to a network and a social product than a pure entertainment product.

34:59 Yeah, >> I'm sure Fortnite despite being the most beloved game in the world 5 years ago like it's not really discussed as much. So, I think there's this holy grail, what you call, John, a neverending game or a forever game, but it's so hard to get there. And maybe with these technologies, they'll have a better shot. >> Yeah. >> Yeah. So, um, in my conversations with game developers, I would say in general, they're all using AI already.

35:22 I think it's AI has become table stakes in terms of just like tools that you use to basically be more productive, right? consider using AI to generate game design docs that generate concept art quickly iterating on on an idea and go to prototype as I mentioned earlier is like shortening that that the distance and time between an initial idea and prototype.

35:43 I think everyone's like universally like really excited about that. Um I think I think of that as sort of like wave one of AI adoption among sort of game studios. Mhm. >> I think the next wave, which is what um Anish has to that's even more excited exciting is how do you make AI to create like new types of gameplay like new game mechanics, new game genres that don't exist today.

36:04 >> Yeah. >> But could massively expand the market for games, right? So you're not just making GTA 6 but like faster and cheaper. You're making a new type of game, right, that couldn't exist before. Yeah. And um I think the I think the the article I wrote um uh earlier on the concept of the never ending game, it's you know it's something that I'm personally passionate about.

36:23 I I'd love to see happen. But um the holy grail for an MMO is a game which infinitely sort of adapts itself to what the players want, right? So >> if you land in a beach in in an MMA and then you're supposed to go to the forest because that's how the game starts. But, you know, maybe you really love the ocean and you decide to stay in the beach and you build a ship and you like sail off into the ocean.

36:46 Like the MMO today couldn't like handle that because it's right scripted linearly that expect you to go into the forest and maybe there's a thing that happens that sets you up on the on a quest. But in an adaptive MMA, one that you know maybe is built using like world models for example, like you could literally generate an entire quest line for someone who really likes the ocean and like now there's pirates that are attacking you.

37:10 There's like a nautical, you know, warfare element, etc. Um, >> and then I feel like sometimes these discussions always go into science fiction, but um, you remember in Ender's game, they have the concept of the mind game. >> Mhm. And it's that the game is just adapting to what Ender wants to do in any given moment in time. So, um, and I think it's it's possible that we'll see our own version of the mind game at some point, >> but I think that would be the game that like like you can't stop playing because it it adapts

37:40 itself to whatever you want to do at any given moment in time. >> I mean, perhaps we're playing it at this moment. >> Yeah, exactly. >> The Matrix. >> Yeah. Well, or Yeah. Or Andrew's game itself, right? I don't even think >> the story is that he's in the game. Yeah. >> Yeah. >> Um what what games do you guys have the most amount of hours in? >> Oh, for me, I've played Witcher 3.

38:06 >> I've completed it twice. >> Nice. >> Half completed a bunch of times and I just absolutely love it. The other game that I actually adore is Dune 2000. >> Oh, >> that was my favorite RTS growing up. It's so so good. And actually, somebody had already ported it over to play on Mac and it's as good as it ever was. >> That's awesome. >> So, for RTS fans, if you haven't played that, I highly recommend Dune 2000.

38:29 >> Yeah, >> I am um I probably put thousands of hours into League of Legends and so um I think >> uh some of my best friends and friendships have been made in that game. And so, I think that has to be probably my my most played game. Um and then probably closely behind that is uh Starcraft. Nice. >> So, I've played all of the all of the versions, you know, back to boot war to Starcraft 2 and so on.

38:52 And it's a classic example of a game which the audience for hardcore RTS is just dwindowed over time cuz it's such an such a hard hard genre to play. Like it matters how like what your APM is. Like you can't play competitively without feeling bad about yourself cuz like everyone everyone else is playing is like so much faster and better at the game than you.

39:12 Um, you know, there's not a lot of developers that are trying to make like new and modern RPS games. Um, you know, because the audience size doesn't support it if you just look at it from a pure like market size and analysis, but with AI, like I'm excited to see an explosion in RTS games, right? And I I will be playing as many of them as I can. >> Nice.

39:33 Nice. I'm a I'm a big Crusader Kings 3 fan, which isn't which isn't really RTS, but it is like it's it's really it's like I would I I describe it to people who don't know about it as like the Sims mixed with Age of Empires. It's like cuz you you get like kind of personality management, >> relationships, politics. >> Yeah. Yeah. Yeah. Exactly. But then you also get kind of this over this grand like overworld of like managing your your dynasty and your realm and stuff like that.

40:03 So, it's so so deep. Um, >> wow. >> Yeah. Anyway, to to close us out, um, >> what's a game that each one of you has been thinking about making for years and and now you finally can. >> I mean, I can go quickly. So, I I grew up um running a bulletin board system, which tells you what my age is. Um, so pre- internet people would dial in. And BBS's were famous for having these async multiplayer games because only one person could phone in at a time.

40:33 So you'd play your turn. You'd phone up, connect, play your turns, log off the next day, play your next turns, and see everyone else's turns at the same time. >> Oh, cool. >> So there's a whole genre of amazing um async multiplayer games, including two written by one of the earliest employees at Google. This was when he was still at Stanford. His name is Ait Patel.

40:50 So, he made a game called Baron Realms Elite and a game called Solar Realms Elite. >> Unfortunately, he lost a source code for Solar Realms Elite, which is the better of the two, but he has a game design doc and he haven't he has an .exe. >> So, I for every genre of the models I've tried like, hey, can you recreate the game and has not been able to do it?

41:09 And I like now have a what feels like a faithful recreation running on my computer. Incredible. >> That's a great story. >> Yeah. I think maybe something along a similar vein for me is I'm I'm a big um Civilization player and I think the original Civ 2 is generated some of my like most endearing sort of childhood memories. It's like the 16 bit like Windows version of of Civ that you need a floppy disc in order to play.

41:36 It's it's not playable today anymore, you know. It's a 16- bit executable. Um but I would love to recreate that game on on modern sort of like a modern hardware stack. Nice. Um, and then you know I I've also long thought about like what does Civ look like with agents, >> right? >> Because um the fantasy of the game is that you are supposed to be the ruler of an empire and like >> um but that fantasy sort of breaks down over the course of Sib where you're microing like literally every unit in your empire.

42:10 >> Yeah. M but you know at some point in time you could imagine hey what if there were SB where every unit was a was an agent >> and they had their own sort of like agenda and goals and then you as the ruler of the empire you're sort of setting strategy and then you're trying to influence your subjects to grow in the direction that you want them to right and you can't just have them do exactly what you want you have to kind of like casual and sort of like >> imping I have amazing news for you this is Crusader Kings

42:39 three. This game exists. This game exists and and doesn't use, you know, like Frontier Frontier AI at all. >> Is that true? It's the same game. >> I mean, yeah. Yeah, pretty much. I mean, you do you have to manage like schemes and like, you know, different people's personalities and their opinion of you and like sometimes people won't do what you want because their opinion of you is too low.

43:04 Like there different way like like if you want to you know conquer a realm or conquer you know a different territory there are so many different ways you can do it. Uh you can like murder people who are like ahead of you in succession. Like it's it's super deep. It's like yeah it's it's it's I think like it it actually feels the closest to me like a kind of you know neverending and endless game like you were talking about before in terms of like what will this feel like when it is super deep and super realistic and you

43:37 get like way way way more permutations of choices. So yeah, you should you should definitely check it out. During the pandemic, um I I got COVID in like the end of 2020 and I started teaching myself C and I started making like a like sort of Japanese style dating sim with Unity. >> Popular. >> Yeah. Yeah. Yeah. Well, cuz it was right around like Unity's IPO and so I was just kind of like curious about, you know, >> Unity.

44:06 >> Unity and like how it worked and like how easy it was. It was like I mean I didn't I like it was basically like branching dialogue trees. So it wasn't really like that advanced in terms of programming and I like didn't really get that far. There was like some basic unit like like kind of like managing health and like thirst and stuff like that. But >> I actually think the dating we're going off on a tangent now but because you brought it up I think the dating sims genre is actually right for a >> Yeah.

44:37 Um so there's a genre called autoto games which are basically you know dating sims for women to basically um go on dates of um with like men of different personalities and archetypes and so on. Um and there's a game called Love and Deep Space which is made by a Chinese developer that is blowing up. I mean it's throwing hundreds of millions in in revenue.

44:56 Um, but the challenge is that it's it takes a lot of time and work to create the content because the expansion packs for the games are basically characters, right? There's a new there's a new guy that you can date and then that is an update for the game and they have to build out this like rich backstory like this guy like you know likes cats and like walks in the beach but like is terrified of heights etc.

45:20 And there like you kind of have to discover all of this. >> Um, but there's no reason why they can't be agents in the future. Yeah. Like these are people that >> have their own personalities. They have their board. It's kind of encoded into their, you know, into their like database and they're um reacting to you in real time as you're having conversations with them.

45:38 They know about your day. I I think there's I think the dating s genre is like wide open right now and right for disruption for for AI. >> Yeah. >> Amazing. >> Amazing. Fun fun note to end on. >> Yeah. For sure. >> Fun note to end on. Thank you guys so much for joining. >> Thank you. So much fun. Thanks. >> This was great. That was cool. Yeah. First of many, huh? Yeah. >> Yeah.