18h ago

William Rinehart proposes antitrust safe harbors for AI labs to conduct joint safety evaluations on unreleased models

Joint evaluations would strictly exclude commercial factors like pricing

0
Original post

@TheZvi Clarity on "yup it's allowed" also helps by removing 'anti-trust as excuse' from a group that maybe just didn't want to actually coordinate

11:22 AM · May 25, 2026 View on X