Greenblatt has tried to put numbers on what political will buys. He sketches a spectrum of how hard the world is trying, from Plan D, roughly today’s world, where maybe ten people inside each company are trying to implement safety measures, up to Plan A, a strong international agreement with real enforcement and a slowdown. His tentative estimate: conditional takeover risk falls from ~45% (Plan D) to ~7% (Plan A). An ~84% relative cut, most of it unlocked by political will, partly directly, partly by buying the time for technical safety work to actually be applied
I now think the decrease in risk would be significantly smaller, though it depends a lot on the implementation quality and exact political will (and how robust and focused it is). See here for some updates (though I use different operationalizations consistent with ai 2040: plan a).
Thanks Ryan, I’ll update the section to your current numbers (Plan D → A = ~45%→~19% rather than ~45%→~7%).
You also say that a near-ideal Plan A is “roughly 2x lower risk,” so if I understand your view correctly, a large part of the gap is implementation quality. But for a plan that calls for an international agreement, my guess is that implementation quality is, in large part, a function of the depth of the political will behind it.
I now think the decrease in risk would be significantly smaller, though it depends a lot on the implementation quality and exact political will (and how robust and focused it is). See here for some updates (though I use different operationalizations consistent with ai 2040: plan a).
Thanks Ryan, I’ll update the section to your current numbers (Plan D → A = ~45%→~19% rather than ~45%→~7%).
You also say that a near-ideal Plan A is “roughly 2x lower risk,” so if I understand your view correctly, a large part of the gap is implementation quality. But for a plan that calls for an international agreement, my guess is that implementation quality is, in large part, a function of the depth of the political will behind it.