An open letter on open weights redraws the AI policy fight — the argument moves from whether to release weights to which obligations should attach when you do
A widely-signed letter on open weights has shifted the terms of the safety-policy argument. The question is no longer a binary about whether open release is acceptable, but which specific obligations — copyright policy, training-data disclosure, downstream information — should attach to it.
This is progress, even for people who disagree with where the letter lands. A binary argument about open release could not be resolved because both sides were describing different risk models. An argument about which obligations attach is tractable — it can be settled clause by clause, and it maps onto how the AI Act actually works.
The concrete test arrives in days. The Commission's enforcement powers activating on 2 August will show whether the carve-out as drafted is workable, because it will be the first time anyone has standing to test it.
ComplexDiscovery — Open weights, open questions: the letter that redrew the AI policy fight → · Truth on the Market — Open Models, Closed Minds: AI Policy Keeps Regulating the Wrong Thing →