I resigned from Anthropic today(twitter.com)
682 points by yurivish 1 day ago | 968 comments
tl;dr: Summary not available.
HN Discussion:
  • The claim that AI is uniquely dangerous is hyperbole; other threats like nuclear weapons and climate change are worse
  • ~Applauds the author for acting on principles even while disagreeing with doomsday takeoff scenarios
  • Skepticism of the author's claims lacks imagination; disaster paths are plausible given emerging capabilities
  • Danger comes from the full system (tools, autonomy, agents, credentials), not the model alone
  • ~Even if we agree AI is dangerous, individual/national action is futile due to geopolitical arms race dynamics