For the last few months I've been documenting the errors LLMs have been making in my projects and others - the crimes they have been committing to projects (pun intended).
Now I've turned it into a game to see if you would spot the issue.
Free to play with keyword matching judgement.
Pay is only to add symantic judging.
alexvervloet•55m ago
I also built it because I hate AI screwing up my projects with little, difficult to spot bugs that compound.
I'm fine eating the server costs to host it, but I wanted to add some advanced judgement for those who want it. This passes your answer to claude to judge with extra metadata passed in from the context of the AI code within the project it made the mistake in.
This means the free to play is just keyword matching.