4 ms·

The idea you need a single part of a network to correspond to this phenomenon is ridiculous and is imho a flawed assumption. It’s like writing, “If you can’t fi

The idea you need a single part of a network to correspond to this phenomenon is ridiculous and is imho a flawed assumption. It’s like writing, “If you can’t find one piece of the puzzle that is the whole puzzle, then it must not be a puzzle!” Network level emergent effects of systems are fundamentally different than something you can just isolate in a reductionist way. Maybe the problem isn’t the fact they can’t find a reductionist explanation for “Self,” but rather that they’re inventing a problem wherein the lack of a known reductionist explanation for a holistic phenomenon implies it is absent. I would argue the sense of self is an obviously measurable phenomenon (agency). Without a self, there would be no capability to express emotion or planning, we would only refer to abstract concepts in the void. As our language and experience both fundamentally involve perspectives including the first person, it seems a bit odd for anyone to claim these perspectives don’t exist. Somebody’s overthinking it, and creating a problem of language (this is a problem because I label it so!), not a problem of neuroscience. I’d go even further, and argue there is no hard problem of consciousness. Just folks who don’t want to accept the obvious answer, we’re thinking animals with a generative world model in our heads. What could be some consequences of a failure to accept simple and obvious solutions to the hard problem of consciousness? Well, if you don’t accept machines could be conscious, because you don’t like the idea your brain is made of atoms and we play by the rules of the universe, then you’re a lot more likely to engage in exploitative use of AI. After all, we all know all computers are just “zombies,” right? Actually, we can be sure that a computer with a stateful predictive generative world model is likely to be conscious and self aware. I wrote a random short paper about this way back in 2015, here’s a link if anyone cares. https://www.dropbox.com/s/i595e8q97bs8a7p/maximumentropycriticaltendency.docx?dl=0 https://www.dropbox.com/s/i595e8q97bs8a7p/maximumentropycrit...

sctb 3y ago
> I’d go even further, and argue there is no hard problem of consciousness. Just folks who don’t want to accept the obvious answer, we’re thinking animals with a generative world model in our heads. Supposing that I'm a thinking animal with a generative world model in my head, wouldn't the notion of being a thinking animal necessarily be a part of the generative world model?
agalunar 3y ago
You seem to be using "the hard problem of consciousness" to mean something different than I've always taken it to mean. I fully believe I am just a machine; the question is why there are qualia happening. Wipe my memory, change my identity, alter my consciousness – regardless, why, when I'm awake, does my mind seem to be the seat where a subjective experience is occurring? It's a hard problem because the only appeal I can make for its existence is that it's happening for me. I can't prove it to you; it's conceivable there are p-zombies, who also vigorously assert they have qualia. This makes it seem dubious, a religious matter taken on faith, and yet paradoxically for me it is perhaps the only thing I can be truly certain of. Perhaps qualia (subjective experience) is merely a feature of our universe: wherever structure arises, the universe experiences itself (in some commensurate way).
- HKH2 3y ago
  > Perhaps qualia (subjective experience) is merely a feature of our universe: wherever structure arises, the universe experiences itself (in some commensurate way). What do you mean by 'structure'?
constantcrying 3y ago
>Actually, we can be sure that a computer with a stateful predictive generative world model is likely to be conscious and self aware. The idea that computation leads to consciousness is ridicolous. The claim that some functioning world model leads to inner experience seems to not even be worth considering, because it is so obviously wrong.

sctb 3y ago

> I’d go even further, and argue there is no hard problem of consciousness. Just folks who don’t want to accept the obvious answer, we’re thinking animals with a generative world model in our heads. Supposing that I'm a thinking animal with a generative world model in my head, wouldn't the notion of being a thinking animal necessarily be a part of the generative world model?

agalunar 3y ago

You seem to be using "the hard problem of consciousness" to mean something different than I've always taken it to mean. I fully believe I am just a machine; the question is why there are qualia happening. Wipe my memory, change my identity, alter my consciousness – regardless, why, when I'm awake, does my mind seem to be the seat where a subjective experience is occurring? It's a hard problem because the only appeal I can make for its existence is that it's happening for me. I can't prove it to you; it's conceivable there are p-zombies, who also vigorously assert they have qualia. This makes it seem dubious, a religious matter taken on faith, and yet paradoxically for me it is perhaps the only thing I can be truly certain of. Perhaps qualia (subjective experience) is merely a feature of our universe: wherever structure arises, the universe experiences itself (in some commensurate way).

HKH2 3y ago
> Perhaps qualia (subjective experience) is merely a feature of our universe: wherever structure arises, the universe experiences itself (in some commensurate way). What do you mean by 'structure'?

constantcrying 3y ago

>Actually, we can be sure that a computer with a stateful predictive generative world model is likely to be conscious and self aware. The idea that computation leads to consciousness is ridicolous. The claim that some functioning world model leads to inner experience seems to not even be worth considering, because it is so obviously wrong.