An Anthropic researcher just gave us a peek at self-improving AI

Training AI models with other AI models has become a very popular goal for neolabs — and now, a researcher in Anthropic’s fellows program has given us