Hacker News Reader

Comment by DanielBMarkham | Hacker News Reader

Or a bumblebee, right?

Beats me. I've given it some thought, and I could speculate.

My point was to separate that part out, though. Acknowledging that there must be a scale is the first step. Once we go there, then we can start talking about what goes into producing one. Some folks don't even want to go there (as this article implies)

Being conscious, in my mind, is the ability to observe information and change your mental model of the world based on it. If that definition works, then animals are certainly conscious.

Perhaps sentience is the ability to do that as a community, over time, and about abstract concepts. Monkeys might be able to locate new sticks in their environment, use them to spear fish, and pass that to their offspring. But they couldn't formulate a theory of virtue over several generations. (At least I don't think so) Perhaps that definition is a starting point?

(Just trying to provide you with some kind of answer)

View on HN

> "Being conscious, in my mind, is the ability to observe information and change your mental model of the world based on it."

Neural networks aren't conscious. Consciousness requires that a locus of ego experiences sensations of qualia.

Just because something is intelligent doesn't mean it is conscious.

I tentatively agree, but this may end up being a distinction without a difference, i.e. if something appears conscious because it's intelligent, for all the outside world can say, it is conscious.

I believe strong AI is going to solve this question over the next several decades.

Intelligence can be unaware. Often they call algorithms "artificial intelligence".

Consciousness cannot be unaware.

Given only a limited set of inputs in a limited space, (e.g. r, g, b, a channels in a jpeg image, and a limited set of filters to separate cats from monkeys), modern neural networks are hardly aware, so they're hardly conscious.

Given only a limited set of inputs in a limited space...

I believe you are making the edge case here, simply because that's where our technology is.

The alternate position is that given a rather larger set of inputs, over a rather larger space, an intelligence could appear to outsiders as being as conscious ans sentient as they are, even though what we're talking about "under the hood" is some simulacrum of sentience, not sentience as we humans experience it.

This is the "if a computer plays chess as well as a human, is the computer really playing chess? Or just mechanically moving things around?"

My point is that with some larger solution space, the question doesn't matter anymore.

> "My point is that with some larger solution space, the question doesn't matter anymore."

The question still matters ethically. If they're not aware, should we put their rights over the rights of humans? Probably not, no.

E.g., If AI is seemingly intelligent, but without consciousness, we wouldn't want to give them a right to self defense.

I understand your question, but I do not currently understand how to process your question.

If I look at something and it has quality A, is it any different to me than if I look at another, fake thing, that also has quality A?

If every way I examine it, it is the same, is it not the same? At least to me?

We are entering the realm of the unknown, perhaps the Great Unknown. Here we cannot reason. It's not rational. It's not irrational. It's trans-rational.

Sidebar: I'll note that you mention consciousness. There is a strong reason why I attempted to define consciousness ahead of sentience. The latter is dependent on the former. Both definitions are controversial.

ADD: It occurs to me that some kind of religion and alien contact are intractably linked, both from our side and from the other side. Interesting.

Given the limitations of computation over large inputs and large spaces, the question is, could sentience be a prerequisite of solving those kinds of problems? All we know about machines and computations is the harder the problem the more their computation is aligned with biology.

My view is that even if we want to have such a scale, creating it would require solving the most tenaciously vexing philosophical problems of all time...

Or, maybe we can come up with a general ordering of sentience, such that we can say that a bumblebee is quite definitely below a baboon, but still, ethical questions will involve picking some cutoff—and that will be arbitrary, contentious, and creepy.

It seems somewhat likely to me that if we manage to quantify or understand sentience across species, we would find it unsettling. What if rats, cows, and pigs all turn out to be well past the hump of vivid awareness, such that killing them really is extinguishing a life quite similar to a human's?

Fortunately there's a lot of good vegan recipes out there...

> "What if rats, cows, and pigs all turn out to be well past the hump of vivid awareness, such that killing them really is extinguishing a life quite similar to a human's?"

Or what about the opposite? What if consciousness is very rare, and that just because something acts with complex behavior doesn't mean they are experiencing?

- Perhaps some humans aren't even conscious, being merely symbolic engines running on a large cultural dataset. It would certainly explain a segment of the population that has never made sense to me.

- Perhaps children aren't yet conscious. I know that I don't remember anything before the age of 4. Maybe I wasn't actually aware yet, being only a biological automaton until a flip was switched in my brain.

- Perhaps animals are conscious, but don't experience a joy sensation, only a hunger/need/pain sensation. In that case, nothing we could do would help them.

- Perhaps animals are conscious, but don't experience a pain sensation, only pleasure -- their reaction to pain being merely instinctual, without actually being felt by the animal's consciousness. In that case, nothing we could do would hurt them.

When you're dealing with consciousness there is no way to know. We can't move the locus of our awareness inside another creature to see if it is also experiencing., or in what ways.

This is my fear as well. I think once we finally come up with a sliding scale, most if not all humans are going to score rather low on it.

If you believe that the future of any intelligent lifeform is to evolve into some form of machine intelligence, and it's obvious that machine intelligences can communicate complex ideas in blinding fast speed -- in parallel channels? Plus live tens or hundreds of thousands of years? Compared to that we're like mayflies, popping into existence and back out in a blink of the eye, doing nothing more than buzzing around. Hell, that's not just true of individual humans. Our entire species might look like that.

Want the answer to Fermi's Paradox? There it is. How much time are you going to spend today trying to observe and understand a individual gnat?

Indeed, no way to know. For me, that means we should act with caution and respect. The opposite stance seems very dangerous.

The stance of "caution and respect", as you put it, is not without its perils.

If we accept that animals are conscious, the amount of laws we'd have to put in place to protect them would very likely have lasting detrimental impacts on the quality of life for the world's poorest peoples.

Given that we know humans to be conscious, I think we must err on the side of protecting humans first and foremost, even if that means to the possible detriment of other potentially conscious entities.

Reply on news.ycombinator.com