It’s not infinite! If you take my cherry picked estimate of the computational power of the human brain, you’ll see we’re just one more round of scaling to have matched the human brain, and then we’re sure to have AGI and make our shareholders immense profits! Just one more scaling, bro!
I think you’re reading some arguments I’m not making. The author seems to be of the opinion that even with infinite resources it’s outright impossible to have a computer that thinks or experiences consciousness, which is obviously a philosophical, not practical, argument, and I don’t agree. I’m not saying we should actually try it, or that it’s doable with our current or foreseeable resources.
That being said, I am defining it. I’m saying that even if we assume that it’s utterly impossible to have consciousness any other way, it’s some incredibly unique combination of the things that make us human and literally any deviation whatsoever makes it all fall apart, there still seems to be a possible path to a computer with consciousness via simulation of that particular and special process. That’s my thing you don’t think I can define: a physics simulation of sufficient fidelity to simulate a thing we already know demonstrates consciousness. “There aren’t enough resources in the universe” would block that path, sure, but that’s not very interesting. Lots of other things could block that path, it’s an insane path to an incredibly difficult goal. But saying “artificial consciousness is impossible, period” means something is blocking it, and the idea that it’s some law of physics that is both crucial to consciousness and can’t be simulated is interesting. I’m struggling to imagine how that would be possible, and if it’s a failure of my imagination I’d like to know. The universe doesn’t have to be computable or deterministic to make a simulation that imitates it so well that observations of the real and simulated physics yield indistinguishable results, and at that point I don’t see anything inherently preventing consciousness.
Well, I don’t really need to for what I’m saying, which is that I don’t see any reason a computer is fundamentally incapable of doing whatever it is that humans do to consider ourselves conscious. Practically incapable, maybe, but not by the nature of what it is. Define it how you like, I don’t see why a computer couldn’t pull the same trick in the distant future.
Personally, though, I define it as something that exhibits awareness of a separation between itself and the rest of the world, with a large fuzzy spectrum between “has it” and “doesn’t have it”. Rocks exhibit no awareness whatsoever, they’re just things with no independence. Plants interact with the world but don’t really seem like they’re doing much more than reacting, which doesn’t demonstrate much, if any. Animals do a lot more, and the argument that at least some are conscious is quite plausible. An LLM kinda, sorta, momentarily glances off of this sometimes, enough that you can at least ask the question, but the fact that outside of training it is literally just an unchanging matrix of numbers and that its only “view” into the world is what you choose to give it means that it can’t be aware of itself versus the world, it’s at best aware of itself versus its training set and/or context window, a tiny keyhole to a photograph of something in the world, means it makes a barely discernible blip on the scale, on the level of plants, and even that might be generous. An artificial consciousness would, in my opinion, need to be independent, self-modifying, and show some of the same traits and behaviors that we see in animals up to and including ourselves with regard to interacting with the rest of the world as a self-contained being.
yud-length posts like these that boil down to “nuh-uh” are why we have the “no debate unless it’s amusing debate” rule, and I can see by the downvotes from local and the reactions that you’ve failed to be amusing
ah, I see. Ok. Well, based on all that, you haven’t actually engaged with anything from the post, nor have you said anything non-trivial. Was hoping that you’d at least say something wrong instead.
Philosophically, right, if you allow me infinite resources, right, to do a thing I don’t actually know how to define,
It’s not infinite! If you take my cherry picked estimate of the computational power of the human brain, you’ll see we’re just one more round of scaling to have matched the human brain, and then we’re sure to have AGI
and make our shareholders immense profits! Just one more scaling, bro!money please!
I think you’re reading some arguments I’m not making. The author seems to be of the opinion that even with infinite resources it’s outright impossible to have a computer that thinks or experiences consciousness, which is obviously a philosophical, not practical, argument, and I don’t agree. I’m not saying we should actually try it, or that it’s doable with our current or foreseeable resources.
That being said, I am defining it. I’m saying that even if we assume that it’s utterly impossible to have consciousness any other way, it’s some incredibly unique combination of the things that make us human and literally any deviation whatsoever makes it all fall apart, there still seems to be a possible path to a computer with consciousness via simulation of that particular and special process. That’s my thing you don’t think I can define: a physics simulation of sufficient fidelity to simulate a thing we already know demonstrates consciousness. “There aren’t enough resources in the universe” would block that path, sure, but that’s not very interesting. Lots of other things could block that path, it’s an insane path to an incredibly difficult goal. But saying “artificial consciousness is impossible, period” means something is blocking it, and the idea that it’s some law of physics that is both crucial to consciousness and can’t be simulated is interesting. I’m struggling to imagine how that would be possible, and if it’s a failure of my imagination I’d like to know. The universe doesn’t have to be computable or deterministic to make a simulation that imitates it so well that observations of the real and simulated physics yield indistinguishable results, and at that point I don’t see anything inherently preventing consciousness.
How do you define consciousness?
Well, I don’t really need to for what I’m saying, which is that I don’t see any reason a computer is fundamentally incapable of doing whatever it is that humans do to consider ourselves conscious. Practically incapable, maybe, but not by the nature of what it is. Define it how you like, I don’t see why a computer couldn’t pull the same trick in the distant future.
Personally, though, I define it as something that exhibits awareness of a separation between itself and the rest of the world, with a large fuzzy spectrum between “has it” and “doesn’t have it”. Rocks exhibit no awareness whatsoever, they’re just things with no independence. Plants interact with the world but don’t really seem like they’re doing much more than reacting, which doesn’t demonstrate much, if any. Animals do a lot more, and the argument that at least some are conscious is quite plausible. An LLM kinda, sorta, momentarily glances off of this sometimes, enough that you can at least ask the question, but the fact that outside of training it is literally just an unchanging matrix of numbers and that its only “view” into the world is what you choose to give it means that it can’t be aware of itself versus the world, it’s at best aware of itself versus its training set and/or context window, a tiny keyhole to a photograph of something in the world, means it makes a barely discernible blip on the scale, on the level of plants, and even that might be generous. An artificial consciousness would, in my opinion, need to be independent, self-modifying, and show some of the same traits and behaviors that we see in animals up to and including ourselves with regard to interacting with the rest of the world as a self-contained being.
yud-length posts like these that boil down to “nuh-uh” are why we have the “no debate unless it’s amusing debate” rule, and I can see by the downvotes from local and the reactions that you’ve failed to be amusing
if I were you I’d reconsider this thread
If anything I’m posting conjures that bozo’s name for any reason, I’ll cut my losses, my apologies.
ah, I see. Ok. Well, based on all that, you haven’t actually engaged with anything from the post, nor have you said anything non-trivial. Was hoping that you’d at least say something wrong instead.