Post

Watching AI agents use the terminal makes me think the CLI was always a really bad UX for humans. LLMs remember so many more flags than me, and are far better at combining them, along with pipes, to get the information it needs.

Sam Rose

@samwho.dev

Lest we not forget that the cost of this superpower is that every so often it might brick your machine or lose hours of work. The real goal of mechanistic interpretability should be identifying the Joker feature and allowing me to easily crank it to 11 on colleagues I don’t like.

9/2/26, 1:01 pm0 reposts1 like

No replies yet.