Could a superpersuasive AI get a high speed rail line between SF and LA constructed for under $50B and under 10 years with existing construction technology?

This is kind of a serious question. Such a thing is not impossible by any means, but it is really hard and requires a lot of persuasion. I think we talk too much about "AI could be 1000x more persuasive than the best humans" and not enough about how, concretely, it could achieve outcomes where the primary bottleneck is persuasion. Another factor is that if superpersuasive AI is widely available, it will be used by different interest groups to achieve contradictory outcomes (e.g. you can't get these costs and timelines if the train needs to go through every "major" central valley city, but there will be groups trying to achieve exactly that).

To be clear also: it's not a rhetorical question either. I think there are assumptions about what AI could do and what persuasion could consist in in which it obviously could achieve this. E.g. if AI could create custom light patterns that could convince anyone that Jesus Christ had personally come down to tell them that their mission in life was to do , then plausibly it could achieve this outcome quite easily.

@AdrienLE i think yes.

@AdrienLE Superhuman doesn't imply omnipotence. I'm not sure there's a pattern of lights that would convince most believing Christians that Jesus wants them to build a HSR. Especially in a world where such a light manipulating AI is known to exist!

@AdrienLE Doesn’t need to be as exotic as that