- Score 83
- Official
Research acceleration: The jagged frontier
OpenAI measures how coding agents are accelerating AI research, where their performance remains uneven, and what this signals about recursive self-improvement.
So what
What this event means by reading role—not a longer recap.
- BuilderExpect agent tooling to help most on well-scoped coding tasks and less on open-ended research work, so scope your automation accordingly.
- ResearcherAgent-assisted research gains appear uneven rather than uniform, which matters for judging claims about recursive self-improvement.
- InvestorOpenAI framing uneven agent performance as a frontier signals where capability and safety narratives may shape future product and policy positioning.
Score dimensions
Higher total means read first. Each bar is one factor we use to rank the system pool. How we score