What happens when AI starts building itself?
This title could be clearer and more informative.Try out Clickbait Shieldfor free (5 uses left this month).
Richard Socher has launched Recursive Superintelligence, a San Francisco AI startup that emerged from stealth with $650 million in funding. The company, co-founded with AI researchers including Peter Norvig and Tim Rocktäschel, aims to build a recursively self-improving AI that can autonomously identify its own weaknesses and redesign itself without human involvement. Their approach centers on 'open-endedness' — inspired by biological evolution and techniques like rainbow teaming — where AI systems co-evolve through iterative self-challenge. Socher distinguishes this from simple AI-assisted improvement, emphasizing full automation of ideation, implementation, and validation of research ideas. He pushes back on the 'neolab' label, insisting the company will ship real products within quarters, not years, and frames future compute allocation as one of humanity's most consequential resource decisions.
