Less Wrong is a community blog devoted to refining the art of human rationality. Please visit our About page for more information.

Wei_Dai comments on The Hero With A Thousand Chances - Less Wrong

63 Post author: Eliezer_Yudkowsky 31 July 2009 04:25AM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Comments (158)

You are viewing a single comment's thread. Show more comments above.

Comment author: Wei_Dai 02 August 2009 01:11:29AM 14 points [-]

I'm not sure why it took me this long to realize this, but by the anthropic principle, the Counter-Force is almost certainly not the anthropic principle, but something that really exists in the world, e.g., some kind of intelligent agent, physical force, principle of magic, or rule of simulation.

Consider two worlds that are otherwise identical except that world A has a real Counter-Force, and world B doesn't. Initially, world A has lower measure since it has higher complexity. But as time goes on, the fraction of world A that survives will massively outweigh the fraction of world B that survives. So, both the Hero and Aerhien should conclude that they're almost certainly in world A.

Comment author: Eliezer_Yudkowsky 02 August 2009 01:26:09AM 2 points [-]

I was waiting to see how long it would take for someone to point this out.

And the "answer" - which, yes, I devised in advance - is that the Dust destroys its worlds completely by compressing the probability out of them, whereupon the probability mass ends up in other worlds. Sort of like "mangled worlds" only these are "squeezed worlds" that have the reality-juice squeezed out of them.

However, anyone in a particular world thus destroyed, gets to observe the takeover of the Dust, since the Dust requires some time to actually compress that world out of existence.

Alternatively we can suppose that a majority of all worlds performing summonings are those in Aerhien's situation: no worlds with a "real" Counter-Force, or worlds that permanently destroyed their Dust, are performing summonings in significant measure. For example, the Dust's existence could be integral to the summoning process.

Comment author: Eliezer_Yudkowsky 18 November 2009 05:07:34AM 4 points [-]

Actually, a better answer than the one I originally thought of, is that the Dust is responsible for their world splitting (in a way which causes it to increase in measure). So worlds which permanently defeat the Dust don't increase in measure past that point, worlds in which the Dust takes over don't summon anyone, and hence the majority of measure in worlds that can still perform summonings are those worlds which have survived but not yet defeated the Dust.

Comment author: rwallace 02 August 2009 04:21:39AM 1 point [-]

I pointed it out a few days ago :-)

The Dust's existence being integral to the summoning process sounds plausible - it would help to explain why there isn't a large measure of surviving/victorious worlds engaging in cross universe commerce. Intuitively, that fits with the Dust squeezing probability mass; perhaps it creates something like a comet's tail of evaporating improbability, some of which the summoning spell taps?

Comment author: cousin_it 12 May 2010 12:37:09PM 0 points [-]

Wait, doesn't the same argument prove the existence of a God in our world that keeps rescuing life from disaster?

Comment author: Wei_Dai 12 May 2010 01:29:13PM 2 points [-]

No, because our world hasn't had as many "lucky coincidences" as Aerhien's. It seems to me that we are not seeing more "lucky coincidences" than a typical evolved intelligent species would see, looking back on the history of its world.

Comment author: FeepingCreature 23 July 2013 06:38:39AM 0 points [-]

I think the most we can say is that there hasn't been a disaster in our history that would have required great luck to stop. Our world has nothing like the dust; our destruction is not nearly that intrinsically assured. So whatever coincidence saved us, anthropomorphically, would not look like an Act of Great Luck; it would look like the sort of thing that you could convince yourself in retrospect must have been more probable than it seemed at the time. Long, drawn-out sequences of individually credible chances.