AI shrinks development cycles, reviving genres studios abandoned
GTA6 took twelve years. Two a16z partners argue AI can now compress game prototyping to weeks, letting studios fund niche genres and new IPs rather than sequels.
The video won't play here. Listen to the audio instead:
The argument · tap a timestamp to hear it
Direct engine control, not just instructions, is what changed game development
Jonathan Lai identifies a breakthrough from roughly a month ago: models can now directly manipulate Blender, Unreal, and Unity instead of just telling users what to do. This means someone with no game dev experience can use Astra to write a game design doc, generate concept art, model, rig, import to engine, add sound, and export a playable game. Astra can even screenshot itself for QA, spotting and fixing its own bugs. This accounts for why early prototypes on X were so surprisingly polished.
— Jonathan LaiLong cycles explain why studios won't risk new IP or unfamiliar genres
Jonathan explains that if a AAA game costs a decade and hundreds of millions of dollars to produce, producers become extremely conservative—approving sequels over new IPs, proven genres like first-person shooters over experimental types. Call of Duty is up to around 20 entries. When AI cuts both timelines and budgets, studios will fund unproven teams and new IPs. This unlocks structural innovation in gaming, not just cost savings.
— Jonathan LaiAI drops the barrier to game-making from professional teams to single creators
Jonathan contends that games are uniquely complex multidisciplinary products, requiring simultaneous expertise in coding, art, design, and audio—a far higher barrier than photography or video. AI fills the skill gaps a solo creator lacks, letting someone with just an idea and reference images iterate toward a prototype. He expects many will quit halfway through, but some will fall in love with game development, causing a surge in creator numbers in the industry.
— Jonathan LaiCounter-Strike and League of Legends grew from mods, not original designs
Jonathan points to Counter-Strike as a Half-Life mod, League of Legends inspired by Dota (itself a Warcraft III mod), and Fortnite and PUBG both descended from Arma's battle royale mod—all became billion-dollar franchises. AI now lets any fan of an IP simply tell Astra: ‘make me a Ghost of Tsushima with a voxel building system.’ If it's fun, it could grow into a new genre the way those historical mods did.
— Jonathan LaiFalling development costs will revive expensive, niche genres like RTS
Anish points to genres like RTS, traditionally expensive to produce with poor returns—no one has made a good iPad RTS despite everyone wanting one—now facing possible full revival as dev costs drop. He also mentions ‘profession simulators,’ an extremely niche category he loves. He recently played a Tokyo subway conductor simulator; these games were economically unfeasible before, but with costs now low, even a tiny player base like his can support profitability.
— Anish AcharyaIP creators historically don't profit from mods made from their own work
Jonathan flags a structural problem: historically, IP creators rarely profit from mods of their own work—Dota's original developers made little from League of Legends, and Half-Life's developers probably didn't make much from Counter-Strike. Now if companies embrace AI-driven remixes, mods, and decompilation-recompilation, the business model is still unclear. Only Minecraft is an exception, having natively benefited from mod ecosystems.
— Jonathan LaiThe real opportunity is new mechanics, not faster versions of existing games
Jonathan frames current studio use of AI for concept art, design docs, and prototypes as the ‘first wave’ of adoption—now standard practice. He argues that the truly exciting ‘second wave’ is using AI to create entirely new game mechanics and genres that don't exist today, expanding the total gaming market rather than just making GTA6 faster and cheaper.
— Jonathan LaiThe ultimate MMO would adapt in real-time to each player's preferences
Anish describes his vision of the ‘never-ending game’: today's MMOs are linearly scripted—if the game wants you to go from beach to forest but you prefer the ocean, it has no way to adapt. An adaptive MMO built with world models could generate an entirely new quest line in real-time for ocean lovers—suddenly spawning pirates and naval combat. He likens this to the ‘mind game’ in Ender's Game that adjusts itself to match Ender's current thoughts, suggesting future games might achieve this in reality.
— Anish AcharyaIn their own words · checked verbatim
Grand Theft Auto 6 is finally coming out next month. Yeah. And like, that game has been in production since 2013, 2014. It's like a 12, 13 year development cycle.
Jonathan Lai3:05
the long dev cycles have a suffocating impact on this innovation. Because if you're a producer at a 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
Jonathan Lai6:07
there's a saying in the games industry, the last 20% of the game is 80% of the work
Anish Acharya10:14
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. Now we're talking about like dynamic abilities.
Anish Acharya15:27
Counter-Strike, for example, which is a really big esports, that was a Half-Life mod. League of Legends, which is one of the most successful MOBAs of all time, was inspired by Defense of the Ancients, Dota, which was itself a Warcraft 3 mod.
Jonathan Lai18:31
historically, the original owners of IP have not benefited that much from the mods that have been created on top of their IP
Jonathan Lai27:47
Glossary
- Modding
- Creating add-on content for games by modifying existing code and assets to add features or change mechanics.
- Decompilation
- Extracting source code and assets from released games to enable recreation, porting, or remastering.
- Three.js
- A JavaScript library for rendering 3D graphics in web browsers.
How to listen
Game entrepreneurs, indie developers, and investors tracking how AI transforms content production economics.
Skip 25:47–26:47, where they debate whether agent orchestration interfaces resemble real-time strategy games—low substance value.