So today’s the last day of “free” Fable tokens before it goes into usage based pricing – so I decided to do my usual “snake game” one-shot I do with new models.
https://www.twicetwice.net/snake
Super impressive. Retro UI, sound, mobile support. Also not bad technically (at a high level) – all in one file, no external dependencies, super easy to deploy.
But at 500+ lines and me not being a game programmer, am I able to review this code confidently? If I was to extend this with multi-player support, change the game mechanics, will I still be confident the AI wouldn’t refactor out functionality along the way?
Lets say in the future, where AI is good enough to have free reign, are we comfortable to shift the accountability from the engineer to the AI – or will we always need a human to be ultimately accountable when something goes wrong?
Leave a Reply