If I didn't come across as disagreeing with some of your claims I probably presented my case insufficiently forcefully.

I think Trump winning the election was knowable with better models of humans, though I don't explicitly state as much. I have a section arguing that the Being is less data-constrained than you imply. I mention economic doubling times in the month, as opposed to your implied years-or-so.

I wrote a blog post a little beforehand arguing essentially the opposite: https://osmarks.net/asi/