A collection of random useful things I've learnt. Inspired and borrowing heavily from Simon Willison, collected in nlothian/TIL.
Browse by topic: AI 3 · animals 1 · chess 1
There's a new set of search engines coming, which instead of searching existing data generate (or supplement) it using AI models. …
In Why Can GPT Learn In-Context? Language Models Secretly Perform Gradient Descent as Meta-Optimizers, researchers investigate how large language models (LLMs) can perform tasks such as classification from examples given as the "context". …
There's a whole field of human-animal communications (or animal-animal) that seems pretty interesting. …
FILMNet does frame interpolation to generate feature aware transitions between images. This makes them actually useful. …
Magnus Carlsen was on the Lex Fridman podcast. The podcast is (as usual) way too long, but there were some interesting points that are applicable outside chess. …