Consciousness is not computation

Scott Aaronson (if memory serves right) had an interesting take on this. He framed his thought in terms of the Turing test, but the argument would apply equally well to the mapped iron bar:

In theory, we could pass any Turing test of finite duration (eg an hour or less) and run in a chat room with finite bandwidth with a giant lookup table. Just look up the entire past of the conversation to see what answer should be given. The lookup can be implemented trivially on any Turing machine (and doesn't even need the machine's full power).

Now there's multiple directions you could take this. Here's Scott with one of them:

> Briefly, Searle proposed a thought experiment—the details don’t concern us here—purporting to show that a computer program could pass the Turing Test, even though the program manifestly lacked anything that a reasonable person would call “intelligence” or “understanding.” (Indeed, Searle argues that no physical system can understand anything “purely by virtue of” the computations that it implements.) In response, many critics said that Searle’s argument was deeply misleading, because it implicitly encouraged us to imagine a computer program that was simplistic in its internal operations—something like the giant lookup table described in Section 4.1. And while it was true, the critics went on, that a giant lookup table wouldn’t “truly understand” its responses, that point is also irrelevant. For the giant lookup table is a philosophical fiction anyway: something that can’t even fit in the observable universe! If we instead imagine a compact, efficient computer program passing the Turing Test, then the situation changes drastically. For now, in order to explain how the program can be so compact and efficient, we’ll need to posit that the program includes representations of abstract concepts, capacities for learning and reasoning, and all sorts of other internal furniture that we would expect to find in a mind.

> Personally, I find this response to Searle extremely interesting—since if correct, it suggests that the distinction between polynomial and exponential complexity has metaphysical significance. According to this response, an exponential-sized lookup table that passed the Turing Test would not be sentient (or conscious, intelligent, self-aware, etc.), but a polynomially-bounded program with exactly the same input/output behavior would be sentient. Furthermore, the latter program would be sentient because it was polynomially-bounded.

See 'The Lookup-Table Argument' in https://www.scottaaronson.com/papers/philos.pdf for the rest. It's all very fascinating.

So a tree falling in a forest with noone around to hear doesn't make a sound?

Defining this word must be one of the favorite exercises on HN. Imo, consciousness is just perception: that's the literal meaning of this word in sanskrit and that's the implied meaning in how we use it - unconscious means "cannot perceive anything". When applied to AI, consciousness would mean an array of sensors connected to some memory block where sensory input become 0s and 1s. Traditional types of consciousness register things from outside, but we can think of an internal consciousness that observes internal events. Perhaps, when some of the sensory inputs observe outputs of the processing unit, that's self consciousness. But even self-consciosness doesn't imply intelligence - the latter is about prediction and cause-result relationships.

> the presence of randomness does not necessarily preclude something from being computational in nature

In the case of single events, it does preclude something from being computational in nature, if you take unpredictability as the hallmark of randomness.

Predictability and repeatability are built in to computation and built out of randomness, as it were.

In one stringent definition of randomness (that of algorithmic information theory) it is literally defined by reference to its uncomputability.

> Now, you can find an interpretation of the iron bar that shows that the iron bar casually and deterministically went through those exact same states for a full 1 second. What is the difference between those two situations? Why would the program be conscious but the iron bar would not be?

Let's say you find an interpretation that perfectly matches 1 second of calculation, after searching an exponential bazillion states.

Then you decide to test the next second, to see if it's the real deal.

Oh, it always fails, with certainty 1 - 1 / [exponential bazillion].

Even if you had perfect random noise you could always find an interpretation to XOR with it and "find" any target you're searching for.

Even if you had flat 0000000..., you could always find an interpretation to XOR with it and "find" any target you're searching for.

The problem is that you actually did the computation when you were finding the "interpretation". Whatever value the rock had becomes irrelevant in a search process like this. The rock isn't doing the computation.

Just like in the library of babel, where the book doesn't have the information, the location of the book has the information.