How about a poll instead so that the reasoning isn't circular? Plausibly most of these sources will simultaneously claim themselves to be best

@TheAllMemeingEye If you believe that in 75 years we will need to defer to a poll of a couple hundred people on Manifold about what institution serves as the most reliable source of truth (because truth verification is still an unsolved problem that we argue about), I'd say humanity has bigger problems than the wording of my prediction.

Also, to be clear, "Krantz" is a very distinct category of a poll.

If you think a "Manifold Poll" (circa 2025) would be the best mechanism for resolution, maybe you should add it to the list as an option.

@Krantz I guess what I'm saying is that unless you pre-commit to using a specific externally-verifiable non-circular method to deciding the resolution source, then I have a high credence that you will unilaterally decide that your eponymous krantz method is best and use that to choose itself, which could be completely divorced from reality with near zero "universal acceptance", while a Manifold poll is probably the easiest way to get an rough approximate idea of actual real world acceptance

@TheAllMemeingEye I think that in 2100, Krantz (the market of propositions), if not "universally accepted as the most reliable trusted authority", will be accepted as the most reliable authority by very few people, in a way clear enough that Krantz (the person) would be unlikely to be confused about it.

But: if the market is about universal acceptance of a source of truth, I think this can require multiple methods of the determining the resolution (which must all agree). If everyone believes that Krantz is the best source of truth, then Krantz (the mechanism), Krantz (the person), a Manifold poll, and Wikipedia should all say so. (You only really need the best source of truth when the truth is hard to determine.)

Relatedly, if there is no universally accepted source of truth, will this resolve N/A?

@jcb This is a very well thought out response. Precisely, this would indeed resolve NA if there were any reasonable dispute about it. A paraphrase of this prediction might read something like, "Will society agree on how to game theoretically agree to what facts are true by 2100?". The thing that should bubble up to the top are the instructions for how to do that.

Thanks

@jcb fair enough, it didn't occur to me that what was meant was that all of them at once would be used. As far as actual predictions go, the seemingly very well thought out AI-2027 forecast implied that in the scenario where we survive a merger of the top US AI labs and the US government have control of a near omniscient superintelligence, so my money is on them