mfs getting pascal mugged left & right
Conversation
I don't think it's a pascal's mugging.
Most people don't seem to think the risks are 10^-20 The *most* sceptical I've talked t think they are in the range of .001% and .00001%. And the median ML researcher seems to think they are like 1-10%.
Further: forum.effectivealtruism.org/posts/vYb2qEyq
2
11
Absolutely. Also, S-risk is bad enough that even a tiny chance is concerning.
1
Show replies
Let's actually do the math. A first-generation nuclear weapon maybe kills about 50,000 people.
Since there's only a 1% chance, the expected amount dead is 500.
Any war is likely to kill more than 500 (didn't the actual war killed more than 10,000?)
Correct decision: no war.
1
Show more replies






