Open Thread: July 2010, Part 2

Alicorn

This thread is for the discussion of Less Wrong topics that have not appeared in recent posts. If a discussion gets unwieldy, celebrate by turning it into a top-level post.

July Part 1

This thread is for the discussion of Less Wrong topics that have not appeared in recent posts. If a discussion gets unwieldy, celebrate by turning it into a top-level post.

July Part 1

Since I assume he doesn't want to have existential risk increase, a credible threat is all that's necessary.

Perhaps you weren't aware, but Eliezer has stated that it's rational to not respond to threats of blackmail. See this comment.

(EDIT: I deleted the rest of this comment since it's redundant given what you've written elsewhere in this thread.)

Perhaps you weren't aware, but Eliezer has stated that it's rational to not respond to threats of blackmail.

I don't think he was talking about human beings there. Obviously you don't want a reputation for being susceptable to being successfully blackmailed, but IMHO, maximising expected utilily results in a strategy which is not as simple as never responding to blackmail threats.

11wedrifid16y

This is true, and yes wfg did imply the threat. (Now, analyzing not advocating and after upvoting the parent...) I'll note that wfg was speculating about going ahead and doing it. After he did it (and given that EY doesn't respond to threats speculative:wga should act now based on the Roko incident) it isn't threat. It is then just a historical sequence of events. It wouldn't even be a particularly unique sequence events. Wfg is far from the only person who responded by punishing SIAI in a way EY would expect to increase existential risk. ie. Not donating to SIAI when they otherwise would have.or by updating their p(EY(SIAI) is a(re) crackpot(s)) and sharing that knowledge. The description RationalWiki would be an example.

-1waitingforgodel16y

These are really interesting points. Just in case you haven't seen the developments on the thread, check out the whole thing here. I'm not sure that blackmail is a good name to use when thinking about my commitment, as it has negative connotations and usually implies a non-public, selfish nature. I'm also pretty sure it's irrational to ignore such things when making decisions. Perhaps not in a game theory sense, but absolutely in the practical life-theory sense. As an example, our entire legal system is based on these sorts of credible threats. If EY feels differently I'm not sure what to say except that I think he's being foolish. I see the game theory he's pretending exempts him from considering others reactions to his actions, I just don't think it's rational to completely ignore new causal information. But like I said earlier, I'm not saying he has to do anything, I'm just making sure we all know that an existential risk reduction of 0.0001% via LW censorship won't actually be a reduction of 0.0001%. (and though you deleted the relevant part, I'd also be down to discuss what a sane moderation system should be like.)