Week 2691

(Previous | Next)

I’m quite pleased with my AI-enabled weekly workflow for processing my notes at work. On Friday, GPT-5 found a completely valid pattern that I hadn’t discerned.

Outside the day job, my mind is mostly on financing the new house (which we signed for on Tuesday), but I also finally dusted off Hexmap. And, also finally, I spent some proper offline time glazing porcelin in the atelier.

Writing

1 Note

There is a growing body of research into LLM self-awareness but I am particularly fascinated by the “evaluation awareness” research by Anthropic into the Hawthorne effect in LLMs. (h/t to @futureshape.bsky.social for the connection)