« first day  last day (15 days later) » 

13:40
@Alex Your concern is based upon what I believe is a dubious assumption - the assumption that humans could potentially know enough about physics/chemistry/biology to create an arbitrarily accurate model of the human brain. I’d say it’s possible that the mind is so complex that humans can never fully “understand” it, since it is also the apparatus with which we do the “understanding.”
 
5 hours later…
18:59
@wizzwizz4 "If you understood, you could describe how those neurons lead to the experience of, say, experiencing the sight of a red thing."

But eventually neuroscience will be able to explain exactly how we see red. From the neurons that receive color input, to the neurons that compare them with memory to give that color a name, and finally the neurons that make the decision of saying "I see red".
Even if it's too complex for us to understand, I don't get why it's a philosophy problem.
It's a neuroscience problem. Philosophy is trying to explain why this extra thing (consciousness, qualia, etc.) supposedly exists.
But it doesn't, it's all just computations.
Like, imagine we wrote a very complex program that can learn things. At first, it would learn how to recognize colors, without being able to explain why it can do so. But eventually it will learn about programming and explain exactly how it works.
You'd ask it "describe how those [lines of code] lead to the experience of, say, experiencing the sight of a red thing.", and it would explain you in detail what code is called from when it gets the input to where it gives the output.
Yup.
And that explanation (or, rather, the analogous one for humans) would be the solution to the qualia problem.
Qualia isn't some big ghostly magical concept; it's just a unit of conscious experience.
We know that it exists, because we experience it.
And we know that, when thoughts happen, neurons fire.
We're pretty sure that the neurons are causing the thoughts.
Does the program experience qualia?
But we don't know how the neurons cause the thoughts.
@Alex In your situation, I think so.
If this is the same AI that you've used before, yes.
So why is philosophy necessary to explain how it works? It's all just math and programming
But the solution to its qualia problem will be different to the solution to ours.
19:08
Completely deterministic
Okay, that doesn't follow.
I can even run the program by hand on a piece of paper if I want
That does follow.
You're saying the paper gets this magical qualia too?
Ish.
Not really.
This is the Chinese Room experiment, correct?
19:10
Yes
So if we ask the chinese room what it thinks about qualia, what will it answer?
A lot of people have said a lot about the Chinese Room experiment, and most of it's wrong.
It will be confused about what qualia means, right?
@Alex Depends; the Chinese Room isn't specified well enough to know that.
It could easily be confused about what qualia means.
Just like you are.
Can we taboo "qualia"?
Yes
Why doesn't "It's all just math and programming, completely deterministic" not follow?
the program is deterministic
19:13
No but this program is
Sounds good.
I write a completely determinstic program, and ask it to learn philosophy of the mind
Wouldn't it be confused about what it's all about?
How does the program work?
Define "learn".
you give it an input question, it gives you an answer that makes sense
GPT-3 does that.
It wouldn't be confused about philosophy of the mind; it would just spout off something that sounds like a philosopher.
19:14
GPT-3 can solve complex questions about math or physics?
Nope.
It can perform basic arithmetic, if you put commas every 3 digits.
Otherwise it second-guesses its calculations and outputs a nearly-random string of numbers.
So then GPT-3 doesn't actually learn things
It produces things that look correct at first glance
I agree with the former, but not the latter.
It can write poetry.
Original poetry.
But anyway, defining "learn" well enough to get an AI to do it is an unsolved problem.
Anyway, suppose we give this program a book on computer science, and it learns all of it perfectly. Then we give it a book on philosophy of the mind, and it gets stuck on the first page and asks:
What's this qualia? CS already lets me answer any question I could have about myself. I know exactly how I work, from the input to the processing and finally output. This description of philosophical zombie is a perfect description of what I am. It takes input, processes it and creates output.
So, we're assuming it's basically human-like, then?
Assume a human, except one that can be explained by CS concepts?
19:18
Yes
But it's all deterministic
There's no randomness in how it works, it's all math.
"Deterministic" is a red herring.
Running it twice on the same input + stored memories produces exactly the same output.
Yeah, but that's irrelevant. Even if it produced a different output, that wouldn't change anything important.
All "random" algorithms can be modelled as deterministic ones, just with extra input (the randomness).
Yeah
Anyway, the program asks "What's qualia?", right?
Yeah.
19:21
It's confused about how explaining qualia is different than explaining every single line of code involved in seeing red and outputting "I see red"
to it there's no more than that
Explaining qualia is a subset of that explanation.
But it only used CS, not philosophy
In explaining those lines of code, you've explained qualia, plus some other stuff.
Philosophy isn't a separate magisterium.
It's going to say "I've already solved the problem, I don't see why the qualia problem is still open"
Well, it isn't.
Not for the AI.
For humans, it is.
19:22
But why?
We've got a Google Translated first chapter of the textbook, and no source code.
Humans are just like the program, but with particles rather than cpu instructions
Or you're saying it's too complex to solve in the case of humans?
Not necessarily.
I'm optimistic; we haven't solved it yet.
But even knowing something is solvable doesn't mean you've solved it.
It seems like some people are saying there's a problem physics can't ever solve
Just because many highly-regarded philosophers are very confused doesn't mean there's no reason to be slightly confused.
(Although, there are some problems physics can never solve; it's just that this probably isn't one of them.)
19:27
I think the second paragraph in this answer summarizes very well my question: philosophy.stackexchange.com/a/6981/47267
"all the stuff our brains do much the same way as any computer"
So why is there this extra problem that physics can't solve?
What do the supposed non-zombies have more than humans have?
Nothing?
Is this a trick question?
oh sorry
A label!
I win.
What do the supposed non-zombies have more than zombies have?
37 secs ago, by wizzwizz4
A label!
(I didn't notice your typo.)
But, there is no difference.
19:29
So your answer is we are all zombies
No, my answer is that zombies are no different from us.
It's the same thing, right?
Pretty much.
Except I don't throw away the useful concept of "consciousness".
You're not claiming the zombies have this magical consciousness, right?
I do experience things; my experience being made out of smaller, non-experience doesn't change that.
@Alex Zombies have consciousness, but it's not magical.
Well, they have "mome".
19:30
And by "experience things", it's all just neurons firing to produce behavior, right?
An experience is neurons firing in a certain area of the brain
And by touching things, it's all just the electromagnetic field, right?
yes?
A computer program is just electrons?
Yes
Going to have to disagree with you there.
Not because we anticipate different experience, but because I think it's not worth discarding this view of the world.
An apple is more than a collection of atoms.
19:32
But of course we don't consider "electrostatic field" to be the same thing as "touching", the same way the program doesn't consider "instructions" as "recognizing colors" or whatever
One is a physics concept, the other is associated with a certain input
Apples roll down hills, and make a soft crunching sound when you bite into them; many other arrangements of atoms don't. So there's an important difference.
Yeah, we don't disagree.
If a tree falls in a forest, does it make a sound?
Depends on how you define sound
If it's just a sound wave, yes
Exactly.
If not, it depends if "sound neurons" fire
Why does so much philosophical debate boil down to "our mappings between concepts and words doesn't line up"?
19:34
And you agree that the inverted spectrum thought experiment makes no sense, right?
It's a waste of everybody's time.
@Alex Well, I used to.
That's "red is green", right?
yeah
yet the brain is the same
If the entire brain is the same, there's a contradiction in the problem statement, like with philosophical zombies.
So obviously that's impossible.
Also, you say "An apple is more than a collection of atoms."
What do you mean by that?
In different brains, however… that's a different story.
19:36
It is a collection of atoms, right?
It's composed of atoms.
It is in the set of all possible things composed of atoms.
But it is not equivalent to "collection of atoms".
And that set is every single thing that exists
Consciousness is in that set too
Nope.
Then we disagree
No, we don't.
I'm being facetious.
(Stupid rate limit.) Light isn't atoms, is it?
19:37
Right lol
Does "heat" exist?
But if you include all particles, then the set includes everything
It includes every object.
It doesn't include sound.
Conciousness is a very special way that certain atoms behave.
Well, not even, really.
It isn't?
It's a way that part of the universe behaves, since it needs more than just atoms to work.
19:39
The deterministic AI I described before can be considered conscious, right?
Well, not needs… you could probably implement it with a billard ball computer, but it'd be really slow and probably solar-system sized.
@Alex Yes, by definition.
By definition of consciousness?
I never said "the program is conscious"
By definition of the AI.
ok then
Does "heat" exist?: The word "heat" is a word produced by our brain when it receives inputs from nerves that come from heat receptors.
Woah, woah. No, I mean actual heat.
The thing that burns your hand.
19:42
Yes, it exists. And the reason I say that is that part of my brain tests the memory where I have the definition of "heat" stored, to the part that stores previous experiences, and finds that an experiences matches that.
The same way a program would say it exists
But there's no philosophy here, it's all neuroscience (or for the program, computer science)
Why do you say it exists?
(As in, what makes you put it in the category, not how does your brain work.)
I just do
(Why does it fit in the category of "things that exist"?)
I see that question, and I just answer "yes"
The same thing for the program
What property does heat have that makes it exist?
Basically, what do you mean by "exists"?
(Don't worry; it's a lot easier than "what do you mean by learn?")
19:45
Well I don't know what I mean by exists
I don't think you can give a precise definition
I've got one!
But the brain basically does statistics
And does so badly.
It sees people use "exists", "heat", and has the inputs that happened when all of those were said
19:46
and just through statistics, comes to the conclusion "heat exists"
Not really… humans don't work that way.
It's a bit more complicated
But it's the same idea
I mean, you could design an AI that worked like that, I guess, but that's completely off for humans.
The brain doesn't work with tokens on as low a level as that.
That explanation's too simplified to make predictions from.
Well we can also think
Once the brain produced those statistical conclusions, we can combine them through logical rules
and other complicated things
The brain doesn't run on those, either.
19:49
Why do you say that?
The program I wrote does, right?
The program does if you say it does, but humans don't.
There's no "logical rules" mechanism in the brain.
How do you know?
There isn't a one-to-one mapping between neurons and concepts, and we don't see the regular macroscopic structure that would be required for the logical rules systems to exist without that.
Also, humans would behave differently.
but most of the brain is just doing statistics, right?
it's associating inputs it sees often together
At a high level, yes, we can implement logical rules. But the brain runs on a certain kind of heuristic.
19:51
to learn what words mean
@Alex It's doing a very specific, very rubbish kind of statistics at a lower level than the concept of a word.
yeah
it's like neural networks in AI
Oh, much better than that.
yes, I agree
but it's all computations
The reason I object to that is because it's not useful for understanding.
19:52
Which leads me again to my question: Why is philosophy of the mind useful?
Neuroscience can answer everything
Eventually, yes.
Philosophy of the mind lets you get there sooner, if you're very careful.
Of course, most philosophers aren't.
still, the concept that zombies and non-zombies could be different confuses me
Philosophy of the mind is less useful nowadays, because we can just wait for neuroscience to catch up, but there's still plenty of stuff it can't do yet.
I don't get what the zombies are supposedly lacking
Psychology, for instance; that's basically the same thing as philosophy of the mind.
@Alex Nothing; philosophers who say there's anything are wrong. I thought we'd moved on from that.
19:54
Ok
2 mins ago, by Alex
but it's all computations
This is a semantic stop sign.
I thought you were saying that consciousness isn't the same as computation that produces conscious behaviour
@Alex It isn't.
They behave wildly differently.
Just because one can be made of the other doesn't mean they're the same thing.
You can make consciousness out of computation and computation out of consciousness. (probably, to both)
Here's an analogy. Have you ever done programming?
Yes
Async?
19:56
Yes.
Weird, I can't send the exact same message twice in a row...
You know how you can build synchronous out of asynchronous (atomic flags acting as locks, etc.), and asynchronous out of synchronous (green threads)?
Yes
Yet they behave completely differently.
Differently enough that they're covered in different books.
The difference between consciousness and computation is like that, but so much bigger I couldn't even tell you the order of magnitude.
Yes, and same for everything else:
Physics
Chemistry
Biology
Neuroscience
Psychology
Where's philosophy of the mind?
Which problem isn't completely solved by the others?
Theoretical psychology.
Philosophy of the mind is theoretical psychology, pretty much.

« first day  last day (15 days later) »