>it's very hard to take these kinds of claims seriously in the wake of GPT-3
GPT has certainly done a lot of remarkable things, though i think his argument would be that while it is pretty good at making responses, it still has a poor memory. that might be an easily fixable contingency like the other anon suggests. nevertheless, i think his general approach to this stuff as a pessimist is really novel. compare it to what searle would rather emphasize which is far more vague imo (not to say useless)...
also i think philosophers are important to the extent that there are still problems in philosophy of mind that haven't been figured out by our current sciences yet. presumably we are all shooting for a human-level ai companion. if it is desired that the companion have a unified consciousness, then we would need to solve the hard problem, and learn to implement genuine common-sense understanding. with that said i just discovered that artificial neuroethology is a field the other day, and it seems like another important piece of at least one of these puzzles
how do you do that (add more nodes and try to make it follow the behaviours of different users?), and would it need to be a robot demiurge to be able to achieve it (i mean gpt-3 already sampled from the entire internet, so we have already broken past the sky as the limit i guess)?
honestly i dont really get the whole robot ethics thing. look how much resources it took just to raise something like gpt-3. you will need an immense amount of resources to make a god-like ai. it isn't going to be an accident but rather a long intentional effort. the question of course is why? i dont really see why you would want a centralized robot god. i doubt you would need something sentient even if you wanted to instantiate something like project cybersyn
i didn't mention him because he isn't really looking at things from the perspective of a waifu engineer as much as the others, but luciano floridi is i think one of the few voices of reason in this whole ai ethics thing. his criticism of the prospects for a sapient superintelligence just follows searle, but his conclusions from there are really insightful. he talks about how humans actually end up modifying our environment and purposely structuring our data in order for ai to better operate (i believe he talks about his position in this video https://www.youtube.com/watch?v=b6o_7HeowY8).
really, at least with our current approach to engineering intelligence, the power of artificial intelligence is really dependant on how much we are willing to conform to behaviours that *they* find most manageable (which also reminds me of this medium article https://medium.com/@jamesbridle/something-is-wrong-on-the-internet-c39c471271d2).
as an aside, it is much like the power of capitalism to shape human culture. adorno complains about how making art a commodity eventually degraded its quality, but at the same time we are the ones consuming all this recycled shit. similar thing with youtube algorithms. they wouldn't be as effective if people had better self-control. ai as we have it is just a tool. if it destroys human civilization, it would only be after us collectively welcoming it with open arms at every step of the way. something something dune
(that was a massive tangent, and im not sure if floridi was looking at things this way)
the other side is about when we should treat robots as people, which just seems like a general ethics, though i think kant (with focus on rationality and capacity for self-determination) gave pretty solid criteria (incidentally, the autist had been fantasizing about alien life on other planets and their inclusion in a moral framework centuries ago)