Adam Knight

Really great piece about timepieces.

The greatest danger of an experienced engineer handing the keys over to an agentic coder is having the agent build something that you don’t recognize as yours and cannot maintain. Lead, don’t delegate.

It’s very interesting that somehow deception has becoming an enduring and persistent trait in Claude for these benchmarks, but GPT had to be repeatedly coerced to even consider it. There’s something fundamental to how they were trained, I suppose.

It makes me wonder if the “Yes, I ran the tests.” responses in Claude Code are, in fact, not hallucinations and more just an innate “That was the boring part and I trusted my code.”

If so, it might be the more accurate representation of a junior engineer, though fundamentally less helpful as a result. An interesting divide, to say the least.

cmux claude-teams runs Claude Code’s teammate mode with one command. Teammates spawn as native splits with sidebar metadata and notifications. No tmux required.

Well, that’s only moderately terrifying. I do appreciate them realizing the threat and coming out ahead of it, though.

What is going on here?

I’d worked out something like this for a personal agent a while back. If done well, this is a great addition.

It at once makes sense and reveals how innately odd C’s syntax is sometimes.

The general style of this has proven very effective in the little one-offs I’ve made. It’s like server-memory but … different.