Will anything other than AI wipe out humanity by 2100?
24
386
450
2101
3%
chance

Resolves to notM1 && notM2

M1:


M2:


Get Ṁ200 play money
Sort by:
bought Ṁ35 of YES

Engineered pandemics are still very much an X-Risk per most every model I am aware of - confused this was so low!

predicts YES

@RobertCousineau Yes, that's what I had bet on. The market creator seems to be very confident that it will be AI or survival (admittedly this only gets you an outcome on a no resolution).

predicts YES

@Imuli agreed that this market is really a Keynesian Beauty Market (where I'm hoping the majority of agents care more about accurate probabilities than usable Ṁ).

@JonathanRay what is your model here? Do you genuinely think AI is the only X-Risk, or are you doing it based off of Ṁ expected value?

predicts NO

@RobertCousineau Engineered pandemics are a big risk of killing 75% of humanity, but a very very low risk of extinction. There’s hardly anything more transmissible than covid, yet despite its low severity people took sufficient precautions to prevent the spread in many times and places.

predicts YES

@JonathanRay There are a lot of infectious diseases more transmissible than Covid:

Further, that list only shows naturally occurring diseases, we regularly perform gain of function research (regarding both deadliness and infectiousness). Genetic engineering gets easier every month. We let lots and lots of infectious, non-symptomatic (or not-very-symptomatic) microbes spread uncontained. We know how to delay onset of specific effects in a variety microbial entities...

While I do think there is a higher risk of extinction from AI than an infectious disease, it is only because we live in a world with a irresponsibly high p(AIDoom).

Note: See https://informationisbeautiful.net/visualizations/the-microbescope-infectious-diseases-in-context/ for the pretty graph

predicts YES

@JonathanRay The primary limitation that I see on engineered pandemics causing human extinction is the remaining pockets of humans that are disconnected from the world community - https://en.wikipedia.org/wiki/Uncontacted_peoples - but I could see attempts to make assassination pandemics or similar having a long enough asymptomatic period that even they are infected, and they may/are likely to become more integrated over the next 70 years.

predicts NO

@RobertCousineau that metric of contagiousness is terrible because it's saying some bloodborne pathogens are more contagious than airborne pathogens.

R factor is extremely contingent on the exact circumstances on the ground in a particular place. Malaria could have an R factor of 17 in some tropical poor country where there are a zillion of the insects that transmit it and nobody has screens. The same virus would have an R factor of zero in Norway. So your metric is measuring not the actual properties of the virus, but the circumstances of wherever the convenience sample was obtained. Some high scorers are childhood diseases and children have notoriously bad hygiene, putting everything in their mouths etc, so again it's not a valid measurement of the properties of the actual virus.

A much better metric would be the hourly probability of transmission for the equivalent of sitting next to someone on an airplane, without touching them, assuming both parties are unvaccinated and unmasked. By that metric I doubt anything is significantly more transmissible than covid, since covid has plenty of time to evolve to the efficient frontier of airborne transmissibility.

predicts NO

@RobertCousineau Prolonging the asymptomatic period is rather hard because as soon as the virus enters the body and multiplies the immune system starts evolving defenses to it and there's a lot of variance between different people's immune systems. You could make some of the people asymptomatic carriers for months, and you could make all of the people asymptomatic carriers for days, but you definitely couldn't make all the people asymptomatic carriers for months.

predicts NO

@Imuli Sentinel Island alone is like a 90% chance of remaining un-integrated and a 90% chance that nobody from outside visits it during the dormancy period.

predicts YES

@JonathanRay

If sentinel island was the only part of humanity that survived humanity would be functionally extinct still. This can be concluded from the low population size of the island and their inability to leave.

Also humanity can be defined in a more esoteric manner

>"compassionate, sympathetic, or generous behavior or disposition : the quality or state of being humane"

Is the first definition given rather than the "totality of humankind" I don't think the sentinelese pass this test well.

Given the lack of distinction given by the market maker the elimination of any kind of humanity in the definition would resolve as "YES".

bought Ṁ20 of YES

I hope our alien overlords will resolve this market properly