Hacker News Reader

Comment by version_five | Hacker News Reader

We'd have to start with understanding what conciousness is. As long as "AI" is a math equation you could write on a big piece of paper if you wanted to, it's very safely in the not conscious camp. Even if we don't know what conciousness is, we can safely rule out many things that it isn't.

Edit: I know about the "it must be an equation" argument, I find it incredibly weak without producing the equation and explaining the mechanism of how it translates into qualitative experience. Saying "it must be so" isn't an argument. That's why I began with saying we'd have to understand what consciousness is in order to consider testing for it. Anyway, I understand how internet discussions go, enjoy

View on HN

A brain, just like any other physical system, can be described with math equations. So I guess we can safely conclude it's not conscious.

EDIT: Ironically, someone just posted this: https://news.ycombinator.com/item?id=34843094

>A brain, just like any other physical system, can be described with math equations.

What makes you so certain? There's a whole literature in both Philosophy and Neuroscience dedicated to the Hard Problem of Consciousness, and no one is even close to figure things out.

You don't think a brain is a physical system? Nothing we know about it indicates otherwise. Large ML models, which are clearly physical systems, behave increasingly like brains. I won't be surprised if GPT-4 or GPT-5 is convincing enough to be treated as a conscious entity by majority of people who interact with it. And when that happens, how can anyone prove that it is not?

Is it?

How do we know our own 'processes' could not be modeled that way?

I think this is the point of the Turing test. At some point, you can't tell if a system is "thinking" or just crunching numbers. And it doesn't matter.

Not convinced that is true, if the universe can be described mathematically then maybe human intelligence can also be written down as some equation on a sufficiently big piece of paper.

Not to handwave too hard here; but do we know, _for certain_, that the universe can be in total described mathematically?

Yes, I know that what I'm hinting at is not falsifiable, burden of proof, etc.

But I think it's a little arrogant of us as a species (nb, I'm not talking to you, sgillen, in particular) to think that since we've come as far as we have, that the mysteries of consciousness and life and the universe and everything are going to be laid bare Real Soon Now.

Do you have any reasons to suspect the universe _cannot_ be described mathematically? Mathematics, unlike physics, is pretty flexible.

We are essentially asking whether the universe could contain a complete encoding of its workings.

The answer is certainly not obvious to me, especially when we have results such as Goedel's theorems.

Goedel’s theorem is not a problem because math is not limited to models of our universe.

I don’t think it’s even controversial that it cannot be fully described mathematically. I don’t think I’ve heard of an epistemological system that would support this. E.g try to reduce the experience qualia down mathematically.

But there’s also no reason this is required for consciousness to be emergent from non-human-brain systems.

I think one of the hard parts of theory of mind is that we probably will never understand the experience of being in things sufficiently different from ourselves. Any computer system will probably be sufficiently different, so we can’t know what it’s like to be an AI.

What’s it like to be a chatbot?

Contemporary understanding and discussion of “what consciousness is” has been so influenced and warped by mechanistic reductionist scientism that it no longer has a viable workable model of consciousness.

Every living, actually conscious entity has four modes of consciousness: waking, dreaming, deep sleep and unconditioned awareness, the substrate of the other three. This knowledge, coming from the Upaniṣads, is intuitively obvious to the unbiased observer; but people today are far from unbiased.

Consciousness cannot be separated from life. And life cannot be manufactured in any laboratory. Yes, technology can imitate and abstract certain functions of living organisms, but consciousness is not, and never will be one of them: because consciousness, or more precisely unconditioned awareness, is the Absolute.

I think the answer lies somewhere here: https://en.wikipedia.org/wiki/Quantum_mind

Reply on news.ycombinator.com