COI is the hard part of neutral evals; nonprofit spinoff or Hugging Face floated

tenobrus · x · 2026-09-13

In the thread on independent third-party model evaluations, tenobrus argues the core difficulty is conflict of interest and establishing a credibly neutral position. He sees paths forward: a non-profit spinoff or a carefully sequestered sub-organization to house the evaluation work.

He adds that Hugging Face doing it itself "is also a great idea." The discussion responds to concerns about the lack of neutral model benchmarking — essentially who can become the industry's Consumer Reports.

Related event: Community Debates Path to Independent Third-Party AI Eval(3 posts)→

Original post →

More from Research

Research channel →