The Race to Recursive Self-Improving AI

The fundamental idea here is to build AI that improves itself, and I think we're getting there. You can think of RSI at many different levels of the stack.

It goes beyond the title with direct discussion of like, think, it's, including: For the first time, we could swap the optimizer as the optimizee, and this leads to a reinvestment of cognitive improvement into itself.

1. 01The fundamental idea here is to build AI that improves itself, and I think we're getting there
2. 02Prompts, harness, weights, and you can go all the way to basically chips
3. 03Our approach to RSI is that we have a multi-agent system working together that's been recursively improved