Richard Socher discusses his vision for a Eureka Machine — a superintelligence capable of inventing most things for humanity. He argues for techno-optimism while cautioning against regulating intelligence itself, advocating for application-specific regulation instead. He introduces Recursive, a company focused on recursive self-improvement of AI, and shares early results showing their system outperforming human experts in model optimization and kernel design.
Summarized by Podsumo
Richard Socher introduces the Eureka Machine — a superintelligence that can invent everything for humanity, starting with scientific discoveries in physics, chemistry, and biology.
Early results from Recursive show AI outperforming humans in kernel optimization, nanoGPT, and nanoChat within days, demonstrating the potential of recursive self-improvement.
Socher argues for regulating specific applications of AI (e.g., surgical robots, autonomous driving) rather than intelligence itself, comparing it to regulating the internet rather than hard drives.
"The Eureka machine is the ultimate invention that will afterwards invent most everything for humanity."
— Richard Socher
"Constitutional AI is fake. It clearly isn’t being adhered to. It was mostly marketing."
— Richard Socher
"If you try to regulate intelligence, it's trying to regulate thought. And that's ridiculous and it's crazy."
— Richard Socher