My argument isn't abstract. Neural networks really are just numerical functions which can be expanded into their equivalent graph representations.
Our output ( mechanical and vibrations ) is also fully quantifiable, thus numbers.
One giant lookup table.
In this case it's a good description, because it correlates with the GPT in reality quite well. But they are not the same thing.
At any rate, this is an ages-old discussion in philosophy, so most likely we are not going to settle this in a Hacker News thread.
If you wanted to focus on thoughts alone, you might want to skip few layers/systems, to give input directly to whatever causes thoughts to happen.
All particles and their interactions could also be represented as numbers. But it just depends on what level we do this, and at what level what kind of complex logic is required.
I'm suggesting that for any given human/environment pair, there is a lookup table that produces that person's actual behaviour in that situation. Modern physics lets us approximate this lookup table, and presumably better physics would give us a better lookup table.
Since human behaviour can in principle be described with a lookup table, I see this as a bad reason to rule out a system as "thinking".
Perhaps there is another way to describe neural nets, one that does not use the language of lookup tables, that makes it feel more like thinking and less like lookups.
One such approach I've seen is looking for embedded world models in neural nets.