Blogger skewers AI labs' anti-distillation stance as the narrowest selective consequentialism

1a3orn · x · 2026-11-21

1a3orn criticizes AI labs' arguments against model distillation as self-serving: labs "peeled knowledge about every job on Earth from everything ever written" and plan to disrupt everyone's jobs, yet claim they themselves shouldn't be copied from. He calls it the weakest form of selective consequentialism, resting on uncited claims that "our safeguards keep AIs safe" and "we deploy safeguards for the public good," and says labs should at least admit the stance conveniently earns them billions at everyone else's expense.

Original post →

More from AGI Musings

AGI Musings channel →