Pink elephant problem: Don't think about a pink elephant.

OK. Now, what are you thinking about? Pink elephants.

Same problem applies to LLMs.

That's OpenCode. The model is Claude Opus, which is probably RL'ed pretty heavily to work with Claude Code. So it's a little less surprising to see it bungle the intentions since it's running in another harness. Still laughable though.

RL - reinforcement learning

I do love the former, but it's been nice to take a break from that and work at a higher level of abstraction.

A clear example of the No True Tinnitus fallacy.

> Tinnitus is a condition when a person hears a ringing sound or a different variety of sounds when no corresponding external sound is present and that other people cannot hear.

https://en.wikipedia.org/wiki/Tinnitus