https://douseful.eu.org/en/tags/ai/Recent content in AI on Leobloghttps://douseful.eu.org/share.pnghttps://douseful.eu.org/share.pngHugoen-USWed, 28 Oct 2026 00:00:00 +0800https://douseful.eu.org/en/translations/bitter-lesson/Wed, 28 Oct 2026 00:00:00 +0800https://douseful.eu.org/en/translations/bitter-lesson/A Chinese translation of Rich Sutton’s 2019 essay. The original is on the author’s site.https://douseful.eu.org/en/notes/clever/Mon, 26 Oct 2026 20:54:05 +0000https://douseful.eu.org/en/notes/clever/The blackmail test in Anthropic’s “A global workspace in language models”: “fake” and “fictional” show up early in the model’s J-space.https://douseful.eu.org/en/notes/this-world-is-not-real/Mon, 26 Oct 2026 20:18:28 +0000https://douseful.eu.org/en/notes/this-world-is-not-real/Anthropic’s research on injecting thoughts into a model’s J-space raised two questions for me: what if the subject were a person, and if the model can feel pain, is this abuse? It also brought to mind the idea planted in Inception.https://douseful.eu.org/en/notes/translating-claudes-thoughts/Mon, 26 Oct 2026 20:04:43 +0000https://douseful.eu.org/en/notes/translating-claudes-thoughts/Thoughts on the video “Translating Claude’s thoughts into language”: train a model to turn activations into text, then turn the text back into activations to check it.