i would read that write-up!

damn this is old. how relevant is this today?

From my reading, almost all of it.

All of it.

that response is not comforting

Microsoft's rwlock implementation was borked up until sometime last year iirc. this stuff is difficult to do correctly

some models are capable of metacognition. i've seen Anthropic's research replicated.

Can you elaborate on what you mean by metacognition and where you’ve seen it in Anthropic’s models?

well, first the model has to actually follow the instructions in agents.md =)

Treating prisoners humanely and preparing them to resume a productive life after prison helps us all.

see also: https://www.reddit.com/r/LocalLLaMA/comments/1qk9vmv/1833x_f...

simple lil setup with qdrant