Is there some sort of rule where ppl who want to deny the real advancements in AI recently are forced to adopt arguments that, if taken seriously (and they don't actually take them seriously), would classify humans as non-sentient?
Current AI is drastically below humans. You don't have to do this.
Humans bullshit all the time. We hallucinate in our sleep and in rare cases while awake.
Current AIs are far below human intellect b/c they're not powerful and efficient enough at the moment to simulate more than a very small fragment of our intelligent. Not b/c they get things wrong sometimes.
Oh, and we know how to adversarially provoke small hallucinations in human beings at will using optical illusions.
Not particularly impressive hallucinations, of course, but still completely false imagery that does not correspond to reality.
as impressive as some illusions can be I think it’s also incredibly impressive how well humans generally across our whole perceived intelligence spectrum can rebuild the world from our senses in ways that we aren’t close to approximating in robotics.
In some ways the controversy doesn’t surprise me, both the quasi-religious over-hype and the opposing recoil from it. AI’s foibles are reminiscent of our own in a very “uncanny-valley” sort of way. The current period is making us reflect on what being human really is, and maybe we’re not sure!
I suppose we should've anticipated that the sense of existential dread that drove a bunch of LWers into bizarre conclusions would arise in the general populace, too, in the aftermath of a huge AI breakthrough.
Exactly, and even if your current position is that it isn’t there NOW, we are forced to confront that it mostly certainly WILL BE, eventually. And I don’t think anyone alive now seriously understands what that means and what it will look like.
