Recursive Superintelligence raises $650M for self-improving AI
Original: What happens when AI starts building itself?
Why This Matters
Represents a major funding milestone for recursive AI research and autonomous system development.
Recursive Superintelligence, a San Francisco AI startup co-founded by Richard Socher, emerged from stealth with $650 million in funding. The company aims to create recursively self-improving AI models that can autonomously identify weaknesses and redesign themselves without human involvement.
The startup is co-founded by Richard Socher, known for founding You.com and work on ImageNet, alongside prominent AI researchers including Peter Norvig and Tim Shi. The team is working toward what Socher calls 'truly recursive, self-improving superintelligence at scale' that would automate the entire process of ideation, implementation, and validation of research ideas. Their approach uses 'open-endedness' - a technical concept where systems can continuously evolve and adapt, similar to biological evolution. Co-founder Tim Rocktäschel previously led open-endedness teams at Google DeepMind and worked on the Genie 3 world model. The company differentiates itself from simple AI improvement by focusing on systems that develop self-awareness of their own shortcomings.