The Race to Recursive Self-Improving AI

MakerMaker.AI claims early empirical signs of true recursive self-improvement, but insists takeoff isn't guaranteed and is extremely hard to verify.

• Defining true RSI: MakerMaker.AI founders distinguish genuine recursive self-improvement — where the model itself is both optimizer and optimizee — from merely using AI tools to help humans build AI, which they call "autocatalysis."
• Numerator vs. denominator framework: They frame capability growth as performance (numerator) divided by resource cost (denominator); true RSI requires both improving simultaneously, not just more capital or compute poured in.

Themes, sentiment and reasons are AI-written paraphrases, checked so they never quote an episode. They may contain errors. Not investment advice. Episodes remain the property of their creators.