[scottmsul]

Something that's always bothered me about many-worlds is that it states the number of worlds should always be increasing exponentially. Which means from an anthropic point of view, the most likely time to be born is the last ever human, with insanely exponential odds. But many people have been born in my lifetime, ergo many-worlds seems unlikely.

[consilient]

The many-worlds interpretation is badly named: the "worlds" are just particularly classical-looking regions of the wavefunction we've chosen to draw some lines around for our own convenience. They're part of the map, not the territory.

It's true that a decohering system will have exponentially more classical-ish parts over time - but the parts get smaller at the same rate, and the total size of the system remains the same.

[ravel-bar-foo]

This reasoning is fallacious. Assuming the theory, the number of worlds in which humanity doesn't exist or already died off also increases exponentially, as does the number of worlds where you are not the last human. These exponential terms balance out to the result that many worlds cannot affect the probability of conditional events at all, once you take account for the anthropic bias.

[kovezd]

I've been curious to learn theories where only a subset of scenarios beget alternative world, and many of these ultimately merge into more probable branches of reality. Keeping the total number of worlds growing at less than O(n*2) with respect to time.

[sudosysgen]

In the MWI there is an infinite number of worlds (as many as there are real numbers). So the total number of worlds grows at O(1) :)

[sudosysgen]

Well, not quite. You'd actually have an infinity of worlds with the cardinality of real numbers. So it would never increase or decrease, in the same way as there are as many numbers between 0 and 1 as numbers between 0 and 2.

[andrewflnr]

This kind of thing is exactly why you shouldn't take those sorts of anthropic argument too seriously.

[kbelder]

Does that argument hold for all of us?