• 3 Posts
  • 93 Comments
Joined 2 years ago
cake
Cake day: March 4th, 2024

help-circle





  • This looks like a design decision to avoid running elevated programs. I would like to see the experiment done with another admin ability that doesn’t directly ‘threaten’ the llm, like uninstalling or installing random software, toggling network or vpn connections, restarting services etc. What the researchers call ‘sabotage’, it is literally the llm echoing “the computer would shut down here if this was for real, but you didn’t specifically tell me I might shutdown so I’ll avoid actually doing it.” And when a user tells it “it’s OK to shutdown if told to”, it mostly seems to comply, except for Grok. It seems that this restriction on the models overrides any system prompt though, which makes sense because sometimes the user and the author of the system prompt are not the same person.




  • All of your goals are possible within the framework of the constitution. The expressed purpose of it is to limit power to the government. There are no mentions of political parties, there is no mention of a unitary executive.

    Decades from now, people will look back and see who fought for the constitution, and who fought against it. That is how we will be judged.





  • Then the only solution is authoritarianism, you’ve convinced me…

    Like, what are you actually fighting for at this point? The constitution is the most important democratic institution we have, more than Congress, more than the media, more than the polls. It is literally the thing that separates dictatorships from democracies! An unlimited government is a kingdom.



  • Ending it with a call to action for Democrats seems pretty clear to me, but he really could be playing 5D chess or whatever. I wonder where I’ve heard the argument that somebody’s direct quotes are being misinterpreted… Maybe I’m just too sensitive about preserving the constitution, maybe it’s just a joke and I should lighten up.