1. Accusations of hypocrisy & regulatory capture
“We should instead focus on keeping powerful chips out of authoritarian hands, stopping industrial‑scale distillation, and requiring safety testing of all sufficiently capable models, open and closed.” – Nevermark
2. Safety testing = de‑facto ban on capable open‑weight models
“If a model fails the test, it should be banned.” – reasonableklout
3. Open‑weight models as a public good vs. dangerous capability
“Open‑weights models that don’t have dangerous capabilities are a public good.” – Nevermark
4. Strategic worry about authoritarian misuse (e.g., China)
“My primary concern is the risk that authoritarian governments—not solely the Chinese Communist Party (CCP), although the CCP is clearly the most capable threat)… build AI models that are more powerful than those built by the US, and use them to achieve permanent military superiority or perpetrate incredibly deep repression of their own people.” – birdsongs
5. Distillation portrayed as hypocritical self‑interest
“The ban on distillation seems hypocritical.” – brcmthrowaway
6. Profit/market motive behind the safety narrative
“Their biggest PR problem is that many people still think of loss‑of‑control/misalignment etc. as sci‑fi.” – reasonableklout
7. Difficulty of defining & enforcing safety tests
“Who runs this test? What happens if this test is too costly or the administrator refuses to allow certain people to participate?” – cogman10
8. Ethical alarm over uncontrolled powerful models
“Because Dario is still thinking in the past… what would someone like Ted Kazniski come up with given the resources of AI.” – zer00eyz