NunoSempere comments on Splitting the timeline as an extinction risk intervention

[NunoSempere]

unless your value function is really weird and considers branches which are too similar to not count as different worlds

It’s unclear to me why this would be so weird