Item — HN Reader

The fact that pushed me over the edge is when I learned a cat twitching in its sleep is dreaming. Apparently all mammals dream.

Now we haven't exactly defined consciousness, but I can't really imagine how you can have dreaming without having an experience that "counts" to me.

Sometimes scientists use "recognizing oneself in a mirror" as the definition of "self-awareness", but I'd be sure to differentiate self-aware and conscious, because infants can't recognize themselves in mirrors, but they almost certainly experience pain that shapes them for life.

2 comments

Philosophers have come up with a long string of traits which supposedly separate us from the animals, and they have a very poor track record:

"Animals can't recognize themselves in a mirror"

"Animals can't use tools"

"Ok, but animals can't make a plan to go get a tool and use it"

"Animals can't make tools" (https://www.youtube.com/watch?v=OYZnsO2ZgWo)

… and so it goes on. My impression at this point is that scientists are spending time patiently disproving these theories only because philosophers are so far off track in suggesting them to begin with.

The article deals with this, slightly tongue in cheek, as:

> A funny pattern in animal behaviour seems to be that whatever difficult task you devise will eventually be done by pigeons

A charitable interpretation is that scientifically, we ascribe the simplest possible behaviour (no conscious thought! No tool making! No tool using!) to animals and treat it as a notable result when these assumptions are proven wrong.

What about a self driving car that, when it is not operational and is just sitting in the garage with its sensors turned off, replays scenarios from the past, creates fictional scenarios loosely or directly based on those experiences, and so on, so as to better organize and refine its data and decision making logic? Gaining additional experience without actually driving on roads.

Basically dreaming, at least in a functional sense. Does that mean it is conscious?

(and what if that same car, when operational and passing a reflective store window, can identify the reflection of itself as being itself, as opposed to being another car?)

What you are describing for the self driving car is completely different.

Programmatically reviewing stored data as defined by some machine learning algorithm is not dreaming.

Unless you are imagining some not yet invented general AI tech to be present in the car.

Why is it not?

How do you know the cat isn't doing something just like that? Because it twitches?

Or are you just saying that because it is "programmatic" that it doesn't count?

Are you honestly arguing that algorithms of the current AI wave may possess consciousness? Or that they are comparable to a cat?

I have no hold ups with programmatic intelligence. I just don't think it has been created yet. Our best attempts are incredibly faint shadows of the general purpose intelligence of living creatures.

There have been some great advancements with targeted AI for specific problems. But these approaches don't involve approaching the world holistically or with any high powered general adaptability. They fail calamitously if the parameters in which they operate are barely tweaked and are no more aware than my microwave.

I have been thinking more and more these days that it may be very difficult for a general intelligence to arise without a comprehensive, coherent world model that is developed over time through interactive, direct experience in the world.

"Are you honestly arguing that algorithms of the current AI wave may possess consciousness? "

I'm arguing that you haven't defined consciousness in a meaningful way, so that question is unanswerable.

It's like debating whether or not a submarine can swim.

I was just responding to your initial claim that a self driving car could dream at night and the suggestion that this might make it conscious.

My claim is that the word "conscious" must be clearly defined before you make any claims about whether or not any entity is conscious.

Same goes for the word "dream."

I must say I find it bizarre that observing an biological entity twitch during a state of decreased activity causes people to speculate that some sort of magic is happening. Or even that they need these concepts that they are unwilling to define to explain it.