Which then gets tossed into a compiler and who knows what kind of code that thing spits out. That's why I only support projects written in assembly by real programmers.
replyIf we accept the “LLM == compiler” cliche (which I don’t, but whatever), then isn’t posting a repo of LLM-produced code equivalent to posting a repo of compiler-produced assembly? Why not just post the prompts and leave it at that? I have about as much interest in reviewing the output of someone’s compiled program as I do looking at the results of their prompt.
replyShame really as it would have been relatively straightforward forward to build in an agent loop that actually tests the games.
reply