Sun on Nostr: There are compelling blog posts and videos arguing that the improvement of AI models ...

There are compelling blog posts and videos arguing that the improvement of AI models has peaked. They use impressive terms like "power law" and mathematics to argue their case. Yet if they were offering to bet on it I would bet against them being right because all these people were wrong about "just add more data won't work". They seem to be aware of this and they always slip in a "there's not more data to add" argument.