Challenges for AI Misuse Prevention: Jurisdictions, Open Models, and Privacy | Dark Hacker News