[notahacker]

My point is also not really about AI, my point is that pointing out that the same arguments that X implies Y could (and have been) applied to virtually every V, W, Z (where X and V/W/Z are both in the same category, in this case the category of "disruptive inventions") and yet Y didn't happen as predicted isn't ad hominem tu quoque fallacy or anything to do with hypocrisy, it's an observation that arguments about the category resulting in Y have tended to be consistently wrong so we probably should treat claims about Y happening because of something else in the category with scepticism...

[satvikpendem]

> where X and V/W/Z are both in the same category, in this case the category of "disruptive inventions"

This is my point, it is not known whether those are all in the same category. The survivorship bias part is that we only know if something were disruptive after the fact, because, well, they disrupted. Therefore, you cannot even compare them like that, never mind the fact that all disruptive technologies are not the same, so you shouldn't be comparing between them anyway.