To qualify, a video game must have its graphics rendered by AI in real time. An AI should be rendering graphics at least 50% of the time in an average playthrough. When the AI is rendering graphics, it should generate at least one image per second. The images should generally depend on the very recent actions of the player - for instance, the player character immediately starts moving left when you press the left button.
It's okay if certain parts of the game are not AI-rendered, like UI elements. However, it doesn't count if the AI is just starting with e.g. a 3D mesh rendering and postprocessing it to make it prettier or fill in frames. It has to generate the video more-or-less from scratch. Specifically, I'd look for a system that generates visual elements like characters and scenery during gameplay and simulates their motion, doing this entirely through ML rather than programmatically, similar to OpenAI's Sora. It's okay if the AI starts with a premade textual or visual prompt, but has to be able to generate and animate new visual elements.
The following is speculation, not resolution criteria: What I'm imagining is a very open-ended game where the AI makes up new landscapes and encounters for you on the fly, sort of like the Mind Game in the book "Ender's Game." This seems doable by a very fast version of OpenAI's Sora, where you just ask it to generate video game footage on-the-fly and conditional on the player's button presses.
I may bet in this market.
Markets with the same resolution criteria
/CDBiddulph/will-a-video-game-with-airendered-g-3b1136cb8e7c
/CDBiddulph/will-a-video-game-with-airendered-g-30dbd8e5eede (this market)
/CDBiddulph/will-a-video-game-with-airendered-g-2e37a190fefe
1,000
3.00@traders This market has resolved, but if you'd like to bet on the number of paid users of AI-rendered video games, please take a look at When will AI-rendered video games have 1 million paid users?
6 years from 2025-2030, all conveniently located in one market!
Resolving YES due to https://oasis-model.github.io
See https://manifold.markets/CDBiddulph/will-a-video-game-with-airendered-g-2e37a190fefe#531sypqnmz3 for discussion
AI-rendered graphics, with quality similar to what we have with current games is probably more GPU intensive than regular rendering. I find it very hard that we could achieve it, with 30-60fps real-time, in such a short frame.
There is also the game logic and controls, that will be another issue.
Might already be possible for very rough AI image generation with low frame rates, not rendered locally.
@GoncaloM Technically only 1 FPS is necessary, but yeah, a very slow framerate might prevent the game from getting enough downloads to close this market
I have personally "made" AI-rendered "games" already, by interfacing with the world model of model-based RL AI DreamerV3. For example: https://www.tumblr.com/multicore-processor/729388032628801536/dreamerv3-playing-its-own-hallucination-of?source=share
But at that level of quality plus needing a pretty good GPU to run it, getting 1 million downloads seems tricky.
@Octopusman388 also with actual tech you should be able to generate graphics for non textual based games based on previous actions of a player, so let's say you manually craft the tutorial and by the end of it you have the graphics for the first chapter and so on (since actually we don't have real time models)
@Octopusman388 Haha, nice try :P The text in AI Dungeon isn't actually being rendered by AI though, and I wouldn't consider the pictures to be generated in real time, which I mentioned was required. I'll edit the description to clarify.
@CDBiddulph the text Is generated by AI, however even if this doesn't count seeing where we're at right now and the speed of progress of the field I wouldn't be surprised if it resolve in yes