What evolves together fits together. We’re using RL to post-train models for Thermo AI research, accelerating algorithmic discovery for our new species of computer. The first sparks of Thermo RSI, in collaboration with @PrimeIntellect.