“Game Design Taste evals” by Raemon

A personal eval I run on new AI models is “design a novel game, with mechanics that have never been seen before.” I do this because this is a domain I have enough taste in that I think I can tell if they are doing something actually creatively interesting. For a long time, they would always say “time manipulation mechanics!” as their first answer. Braid was cool, guys, but, it's almost 20 years old now. But reasoning models started getting novel. Later models started getting reasonably fun. o1 OpenAI's o1 started to impress me. It came up with: Scratch and sniff board game, where what cards smell like determine what powers they have. Okay I'm not sure what that means and it's not obviously functionally different from regular vision-based cards. I think the implementation details it gave didn't actually do much that a regular board game with the powers written down normally would have. When I thought about it myself I did come up with some things you might do with it. But, I sure have never heard of a scratch and sniff boardgame. Opus 4.6 I gave Opus 4.6 a prompt beginning with "Generate a game design [...] --- Outline: (00:40) o1 (01:19) Opus 4.6 (03:00) What Makes It Strategically Deep (04:09) Opus 4.6 Round #2, with less meta bullshit (06:23) Opus 5.5 (08:25) Creative Concepts? (15:15) My guess at your verdict (19:46) How much did this cost? How much subjective time was this? (21:26) Once more, with "no nerd bullshit" but only medium effort (23:23) What was the assignment? (23:48) Ray's chain of thought as he attempts the assignment (26:26) Okay, game concept: Fluidtorio (28:54) Did the AI understand the assignment? (33:31) Do they understand "fun?" (36:33) Can they manipulate and derive novel concepts? (37:57) Generalized Taste The original text contained 2 footnotes which were omitted from this narration. --- First published: October 12th, 2026 Source: https://www.lesswrong.com/posts/f6zuAfteYtupddiqF/game-design-taste-evals --- Narrated by TYPE III AUDIO.