Research updates from the Alignment Research Center on mechanistic estimation, a proposed pipeline for estimating rare catastrophic AI behaviours from the internals of learned networks. Issues cover the matching sampling principle, a white-box estimation contest on random MLPs, and methods for estimating expectations of random products.
Website: https://www.alignment.org/
DYNOMIGHT is a personal blog-style newsletter by Dynomight covering speculative essays on biology, evolution, and AI alignment, often with contrarian takes. Readers get long-form, witty reasoning about why biological systems work the way they do and what evolution's design choices imply for building intelligent systems.
A personal essay blog by Dynomight covering biology, evolution, AI alignment and quirky reflections on internet discourse, with a wry, self-aware voice. Readers get speculative theories and playful philosophical takes on why the body and minds work the way they do.
Dwarkesh Patel's podcast-style newsletter posts long-form interview episodes with AI researchers and historians, with links and sponsor reads. Recent issues cover recursive self-improvement, agent swarms, alignment, and the history of the Spanish conquest of the Aztec and Inca empires.
One short email a week with hand-picked newsletters from the index. Free, no spam, unsubscribe anytime.
