Stuart_Armstrong comments on AI box: AI has one shot at avoiding destruction - what might it say? - Less Wrong
You are viewing a comment permalink. View the original post to see all comments and the full post content.
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (354)
Your mother will die in the next three minutes, at 12 seconds past the minute precisely. Your partner will die in the next year, even if you believe me. However, I can prevent it.
If I can just load up a new copy of you in 5 minutes, I'd say AI DESTROYED but check the prediction.
If I can just flip off the monitor, call mom, and confirm she's dead in a few minutes, then, well, I'd totally be curious enough to cheat. It's not like you can do any harm to me with the monitor turned off for a few minutes, and I can type "AI DESTROYED" with my eyes closed, much less being able to look at the keyboard. I'd probably leave the monitor off until I knew the log was gone, just to be safe.
If I can't cheat, well, I'm not giving you 5 minutes of hacking me, since you clearly want it so badly. AI DESTROYED.
I confess I would have no clue what to do if you were correct ^_^
Kudos on including an immediate and easily-tested prediction. I'd probably tighten up those factors and drop the emotional manipulation.
Transparent and clumsy attempts at emotional manipulation earn a big ol' AI DESTROYED.