What "curiosity" is there in telling Claude: "Do Quake, but Rust this time"?
Taking apart a codebase and rebuilding it in another language can be a deeply fun and rewarding experience if you're trying to understand a system better or sharpen your language skills or engage with an old problem in a new way. But none of that is what's happening here.
Having an LLM do things like this when there's no other practical need for it only accomplishes the cheap dopamine hit of "result", and people who do that are only mortgaging their own curiosity and increasing their dependency on a machine to run their own mind for them. Maybe I'm just a curmudgeon, but that seems pretty worthy of 'negativity' to me.
They said they worked on it over about 4 months. I'm assuming that takes a pretty decent amount of prompting and therefore effort and learning.