I feel like I need to step in here because people are once again getting confused about measure, identity, and decision theory in ways I thought we cleared up circa 2008-2009.
First: The whole “measure declining by choice” framing is confused. You’re not “spending” measure like some kind of quantum currency. The measure *describes* the Born probabilities; it’s not something you optimize for directly any more than you should optimize for having higher probabilities in your belief distribution.
Second: The apparent “splitting” of worlds isn’t fundamentally different between quantum events, daily choices, and life-changing decisions. It’s all part of the same unified wavefunction evolving according to the same physics. The distinction being drawn here is anthropocentric and not particularly meaningful from the perspective of quantum mechanics.
What *is* relevant is how you handle subjective anticipation of future experiences. But note that “caring about measure” in the way described would lead to obviously wrong decisions—like refusing to make any choices at all to “preserve measure,” which would itself be a choice (!).
If you’re actually trying to maximize expected utility across the multiverse (which is what you should be doing), then the Born probabilities handle everything correctly without need for additional complexity. The framework I laid out in Quantum Ethics handles this cleanly.
And please, can we stop with the quantum suicide thought experiments? They’re actively harmful to clear thinking about decision theory and anthropics. I literally wrote “Don’t Un-think the Quantum” to address exactly these kinds of confusions.
(Though I suppose I should be somewhat grateful that at least nobody in this thread has brought up p-zombies or consciousness crystals yet...)
[Edit: To be clear, this isn’t meant to discourage exploration of these ideas. But we should build on existing work rather than repeatedly discovering the same confusions.]
RationalSkeptic · 1h > like refusing to make any choices at all to “preserve measure,”
This made me laugh out loud. Talk about Pascal’s Mugging via quantum mechanics...
Eli · 45m Indeed. Though I’d note that proper handling of Pascal’s Mugging itself requires getting anthropics right first...
Eli · 2h
*sigh*
I feel like I need to step in here because people are once again getting confused about measure, identity, and decision theory in ways I thought we cleared up circa 2008-2009.
First: The whole “measure declining by choice” framing is confused. You’re not “spending” measure like some kind of quantum currency. The measure *describes* the Born probabilities; it’s not something you optimize for directly any more than you should optimize for having higher probabilities in your belief distribution.
Second: The apparent “splitting” of worlds isn’t fundamentally different between quantum events, daily choices, and life-changing decisions. It’s all part of the same unified wavefunction evolving according to the same physics. The distinction being drawn here is anthropocentric and not particularly meaningful from the perspective of quantum mechanics.
What *is* relevant is how you handle subjective anticipation of future experiences. But note that “caring about measure” in the way described would lead to obviously wrong decisions—like refusing to make any choices at all to “preserve measure,” which would itself be a choice (!).
If you’re actually trying to maximize expected utility across the multiverse (which is what you should be doing), then the Born probabilities handle everything correctly without need for additional complexity. The framework I laid out in Quantum Ethics handles this cleanly.
And please, can we stop with the quantum suicide thought experiments? They’re actively harmful to clear thinking about decision theory and anthropics. I literally wrote “Don’t Un-think the Quantum” to address exactly these kinds of confusions.
(Though I suppose I should be somewhat grateful that at least nobody in this thread has brought up p-zombies or consciousness crystals yet...)
[Edit: To be clear, this isn’t meant to discourage exploration of these ideas. But we should build on existing work rather than repeatedly discovering the same confusions.]
RationalSkeptic · 1h
> like refusing to make any choices at all to “preserve measure,”
This made me laugh out loud. Talk about Pascal’s Mugging via quantum mechanics...
Eli · 45m
Indeed. Though I’d note that proper handling of Pascal’s Mugging itself requires getting anthropics right first...