A point I will make next week is that I believe the salient fact of the Hugging Face incident is not any of the discourse about consciousness or agency, but that the internet and all adjacent infrastructure offer a common manipulable/executable substrate. visit.bodleian.ox.ac.uk/event/2026/o...

Show Post

Mechanistic Interpretability is making the rounds again, so I'm re-upping my ask for a short explainer (under 1,000 words) of how topic modeling is supposed to solve AI safety.

Show Post

Went through it but zero mention of an obvious contender: data shape. Latest models like Qwen-3 or Kimi are trained on vast amounts of "clean" synthetic data/environments and this impacts superposition/features.

Show Post

one fun thing about driving through Cambridge is that you can see things like William James’s former house covered in skeletons

Show Post

so basically i took a turn from humanities to ai, all for math to be finally humanities-pilled.

Show Post