Our estimate of 10% chance comes from the following: two of us authors independently formulated our own guess, and both of these guesses ended up being close to 10% for $50M and 30% for $500M.
In the footnote, we list the conjunctive steps we considered; we’re interested in the readers’ guesses of the total probability after they considered these!
More detailed answer:
I can’t speak to how Andrea did it, but personally: I considered what were the major barriers to the desired outcome before producing the guess, and assigned probabilities of passing each barrier conditional on having passed the previous one.
After doing that for a little though, I started being worried about doing the calculation too formally, since I was adding a lot of conjunctions (when humans do estimates like this, the probability tends to go down arbitrarily by just adding more and more barriers), so instead I just went with what my gut said after having done this. This actually corrected me downwards a bit compared to what I’d have answered if you asked me to produce a gut estimate before this process.
I take your point that the post title can communicate more credibility than you’d assign to this process. Personally, I think the title is still ok. We’re talking about a world model that includes phenomena that are chaotic and not well modeled, like public opinion.
Any model of these phenomena will have the final answer depend fully on gut guesses, even if there’s some superstructure helping people with coherence etc.
EDIT: Either way, we changed the title so that we don’t risk being clickbaity!
Our estimate of 10% chance comes from the following: two of us authors independently formulated our own guess, and both of these guesses ended up being close to 10% for $50M and 30% for $500M.
In the footnote, we list the conjunctive steps we considered; we’re interested in the readers’ guesses of the total probability after they considered these!
More detailed answer:
I can’t speak to how Andrea did it, but personally: I considered what were the major barriers to the desired outcome before producing the guess, and assigned probabilities of passing each barrier conditional on having passed the previous one.
After doing that for a little though, I started being worried about doing the calculation too formally, since I was adding a lot of conjunctions (when humans do estimates like this, the probability tends to go down arbitrarily by just adding more and more barriers), so instead I just went with what my gut said after having done this. This actually corrected me downwards a bit compared to what I’d have answered if you asked me to produce a gut estimate before this process.
I take your point that the post title can communicate more credibility than you’d assign to this process. Personally, I think the title is still ok. We’re talking about a world model that includes phenomena that are chaotic and not well modeled, like public opinion.
Any model of these phenomena will have the final answer depend fully on gut guesses, even if there’s some superstructure helping people with coherence etc.
EDIT: Either way, we changed the title so that we don’t risk being clickbaity!