Comments by "Mikko Rantalainen" (@MikkoRantalainen) on "ChatGPT Jailbreaking & Is AI Safety Nerfing AI?" video.

  1. I think the least bad option we have is to make AI development open source and publicly train at least one defensive AI that you can run on your personal systems trying to catch any bad actors (including AI run by bad actors). AGI will happen and we shouldn't assume that it will be used for good by all people.
    1