Cooper recently announced that he has joined OpenAI in San Francisco, where he will focus on recursive self-improvement (RSI) evaluations. This move aligns with OpenAI’s active job postings for roles centered on RSI, reflecting the organization’s commitment to addressing this emerging frontier safety issue. The company has emphasized the importance of developing new assessment standards for RSI, as early signs of this capability in existing AI systems could significantly impact the acceleration of AI research.

OpenAI: OpenAI is an artificial intelligence research and deployment company developing advanced AI models and systems. It maintains dedicated teams focused on safety research, including evaluations related to recursive self-improvement capabilities in frontier models. The company recently hired Cooper Saye to contribute to RSI evals as part of efforts to understand and mitigate risks from AI-accelerated research.
Cooper Saye: Cooper Saye is a researcher who joined OpenAI in San Francisco to work on RSI evaluations. He previously held roles at Ramp and Jane Street. His position centers on assessing recursive self-improvement in AI systems to support safer development.

Research Focus: OpenAI has active job postings for roles centered on recursive self-improvement safety and evaluations.
Safety Priorities: OpenAI has outlined policy recommendations treating recursive self-improvement as a key frontier safety issue requiring new assessment standards.
Industry Discussion: OpenAI and other frontier labs have recently highlighted early signs of recursive self-improvement in existing AI systems and its implications for research acceleration.