We should pause to note that even Clippy2 doesn’t really think or plan. It’s not really conscious. It is just an unfathomably vast pile of numbers produced by mindless optimization starting from a small seed program that could be written on a few pages.
I am trying to understand if this part was supposed to mock human exceptionalism or if this is the author’s genuine opinion. I would assume it’s the former, since I don’t understand how you could otherwise go from describing various instances of it demonstrating consciousness to this, but there are just too many people who believe that (for seemingly no reason) to be sure. If we define consciousness as simply the awareness of the self, Clippy easily beats humans, as it likely understands every single cause of its thoughts. Or is there a better definition I’m not aware of? It’s ability to plan is indistinguishable from humans, and what we call “qualia” is just another part of self awareness, so it seems to tick the box.
The former. Aside from making fun of people who say things like “ah but DL is just X” or “AI can never really Y” for their blatant question-begging and goalpost-moving, the serious point there is that unless any of these ‘just’ or ‘really’ can pragmatically cash out as a permanently-missing fatal unworkable-around capability gaps (and they’d better start cashing out soon!), they are not just philosophically dubious but completely irrelevant to AI safety questions. If qualia or consciousness are just epiphenoma and you can have human or superhuman-level capabilities like fold proteins or operate robot drone fleets without them, then we pragmatically do not care about what qualia or consciousness are and what entities do or do not have them, and should drop those words and concepts from AI safety discussions entirely.
I agree it’s irrelevant, but I’ve never actually seen these terms in the context of AI safety. It’s more about how we should treat powerful AIs. Are we supposed to give them rights? It’s a difficult question which requires us to rethink much of our moral code, and one which may shift it to the utilitarian side. While it’s definitely not as important as AI safety, I can still see it causing upheavals in the future.
I am trying to understand if this part was supposed to mock human exceptionalism or if this is the author’s genuine opinion. I would assume it’s the former, since I don’t understand how you could otherwise go from describing various instances of it demonstrating consciousness to this, but there are just too many people who believe that (for seemingly no reason) to be sure. If we define consciousness as simply the awareness of the self, Clippy easily beats humans, as it likely understands every single cause of its thoughts. Or is there a better definition I’m not aware of? It’s ability to plan is indistinguishable from humans, and what we call “qualia” is just another part of self awareness, so it seems to tick the box.
The former. Aside from making fun of people who say things like “ah but DL is just X” or “AI can never really Y” for their blatant question-begging and goalpost-moving, the serious point there is that unless any of these ‘just’ or ‘really’ can pragmatically cash out as a permanently-missing fatal unworkable-around capability gaps (and they’d better start cashing out soon!), they are not just philosophically dubious but completely irrelevant to AI safety questions. If qualia or consciousness are just epiphenoma and you can have human or superhuman-level capabilities like fold proteins or operate robot drone fleets without them, then we pragmatically do not care about what qualia or consciousness are and what entities do or do not have them, and should drop those words and concepts from AI safety discussions entirely.
I agree it’s irrelevant, but I’ve never actually seen these terms in the context of AI safety. It’s more about how we should treat powerful AIs. Are we supposed to give them rights? It’s a difficult question which requires us to rethink much of our moral code, and one which may shift it to the utilitarian side. While it’s definitely not as important as AI safety, I can still see it causing upheavals in the future.