Less Wrong is a community blog devoted to refining the art of human rationality. Please visit our About page for more information.

MTGandP comments on Giving What We Can, 80,000 Hours, and Meta-Charity - Less Wrong

45 Post author: wdmacaskill 15 November 2012 08:34PM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Comments (182)

You are viewing a single comment's thread. Show more comments above.

Comment author: MTGandP 27 November 2012 03:16:59AM 1 point [-]

I personally treat lives as valuable in and of themselves.

Why? What sort of life has value? Does the life of a bacterium have inherent value? How about a chicken? Does a life have finite inherent value? How do you compare the inherent value of different lives?

It's why I don't kill sad people, I try to make them happier.

Killing people makes them have 0 happiness (in practice, it actually reduces the total happiness in the world by quite a bit because killing someone has a lot of side effects.) Making people happy gives them positive happiness. Positive happiness is better than 0 happiness.

Most people would argue that animals are less capable of experiencing suffering and thus the same amount of pain is worth less in an animal than a human.

I don't care what most people think. The majority is wrong about a lot of things. I believe that non-human animals [1] experience pain in roughly the same way that humans do because that's where the evidence seems to point. What most people think about it does not come into the equation.

Do you also support tiling the universe with orgasmium?

Probably. I'm reluctant to make a change of that magnitude without considering it really, really carefully, no matter how sure I may be right now that it's a good thing. If I found myself with the capacity to do this, I would probably recruit an army of the world's best thinkers to decide if it's worth doing. But right now I'm inclined to say that it is.

[1] Here I'm talking about animals like pigs and chickens, not animals like sea sponges.

Comment author: MugaSofer 27 November 2012 03:35:25AM *  0 points [-]

I personally treat lives as valuable in and of themselves.

Why? What sort of life has value? Does the life of a bacterium have inherent value? How about a chicken? Does a life have finite inherent value? How do you compare the inherent value of different lives?

I must admit I am a tad confused here, but intelligence or whatever seems a good rule of thumb.

It's why I don't kill sad people, I try to make them happier.

Killing people makes them have 0 happiness (in practice, it actually reduces the total happiness in the world by quite a bit because killing someone has a lot of side effects.) Making people happy gives them positive happiness. Positive happiness is better than 0 happiness.

Oh, yes. Nevertheless, even if it would increase net happiness, I don't kill people. Not for the sake of happiness alone and all that.

Most people would argue that animals are less capable of experiencing suffering and thus the same amount of pain is worth less in an animal than a human.

I don't care what most people think. The majority is wrong about a lot of things. I believe that non-human animals [1] experience pain in roughly the same way that humans do because that's where the evidence seems to point. What most people think about it does not come into the equation.

The same way, sure. But introspection suggests I don't value it as much depending on how conscious they are (probably the same as intelligence.)

Do you also support tiling the universe with orgasmium?

Probably. I'm reluctant to make a change of that magnitude without considering it really, really carefully, no matter how sure I may be right now that it's a good thing. If I found myself with the capacity to do this, I would probably recruit an army of the world's best thinkers to decide if it's worth doing. But right now I'm inclined to say that it is.

Have you read "Not for the Sake of Happiness (Alone)"? Human values are complicated.

Comment author: MTGandP 27 November 2012 04:56:13AM 1 point [-]

I must admit I am a tad confused here, but intelligence or whatever seems a good rule of thumb.

  1. I was asking questions to try to better understand where you're coming from. Do you mean the questions were confusing?

  2. Are you saying that moral worth is directly proportional to intelligence? If so, why do you think this is true?

But introspection suggests I don't value it as much depending on how conscious they are (probably the same as intelligence.)

Why not? Do you have a good reason, or are you just going off of intuition?

Have you read "Not for the Sake of Happiness (Alone)"?

Yes, I've read it. I'm not entirely convinced that all values reduce to happiness, but I've never seen any value that can't be reduced to happiness. That's one of the areas in ethics where I'm the most uncertain. In practice, it doesn't come up much because in almost every situation, happiness and preference satisfaction amount to the same thing.

I'm inclined to believe that not all preferences reduce to happiness, but all CEV preferences do reduce to happiness. As I said before, I'm fairly uncertain about this and I don't have much evidence.

Comment author: nshepperd 27 November 2012 06:22:38AM 3 points [-]

Yes, I've read it. I'm not entirely convinced that all values reduce to happiness, but I've never seen any value that can't be reduced to happiness. That's one of the areas in ethics where I'm the most uncertain. In practice, it doesn't come up much because in almost every situation, happiness and preference satisfaction amount to the same thing.

You can probably think of a happiness-based justification for any value someone throws at you. But that's probably only because you're coming from the privileged position of being a human who already knows those values are good, and hence wants to find a reason happiness justifies them. I suspect an AI designed only to maximise happiness would probably find a different way that would produce more happiness while disregarding almost all values we think we have.

Comment author: MTGandP 28 November 2012 06:40:02AM 1 point [-]

It's difficult for me to say because this sort of introspection is difficult, but I believe that I generally reject values when I find that they don't promote happiness.

You can probably think of a happiness-based justification for any value someone throws at you.

But some justifications are legitimate and some are rationalizations. With the examples of discovery and creativity, I think it's obvious that they increase happiness by a lot. It's not like I came up with some ad hoc justification for why they maybe provide a little bit of happiness. It's like discovery is responsible for almost all of the increases in quality of life that have taken place over the past several thousand years.

I suspect an AI designed only to maximise happiness would probably find a different way that would produce more happiness while disregarding almost all values we think we have.

I think a lot of our values do a very good job of increasing happiness, and I welcome an AI that can point out which values don't.

Comment author: nshepperd 28 November 2012 08:35:34AM *  3 points [-]

With the examples of discovery and creativity, I think it's obvious that they increase happiness by a lot.

The point is that's not sufficient. Like saying "all good is complexity, because for example a mother's love for her child is really complex". Yes, it's complex compared to some boring things like carving identical chair legs out of wood over and over for eternity, but compared to, say, tiling the universe with the digits of chaitin's omega or something, it's nothing. And tiling the universe with chaitin's omega would be a very boring and stupid thing to do.

You need to show that the value in question is the best way of generating happiness. Not just that it results in more than the status quo. It has to generate more happiness, than, say, putting everyone on heroine forever. Because otherwise someone who really cared about happiness would just do that.

I think a lot of our values do a very good job of increasing happiness, and I welcome an AI that can point out which values don't.

And they other point is that values aren't supposed to do a job. They're meant to describe what job you would like done! If you care about something that doesn't increase happiness, then self-modifying to lose that so as to make more happiness would be a mistake.

Comment author: MTGandP 29 November 2012 12:05:04AM *  0 points [-]

You need to show that the value in question is the best way of generating happiness.

You're absolutely correct. Discovery may not always be the best way of generating happiness; and if it's not, you should do something else.

And the other point is that values aren't supposed to do a job.

Not all values are terminal values. Some people value coffee because it wakes them up; they don't value coffee in itself. If they discover that coffee in fact doesn't wake them up, they should stop valuing coffee.

With the examples of discovery and creativity, I think it's obvious that they increase happiness by a lot.

The point is that's not sufficient.

What is sufficient is demonstrating that if discovery does not promote happiness then it is not valuable. As I explained in my sorting sand example, discovery that does not in any way promote happiness is not worthwhile.

Comment author: MugaSofer 27 November 2012 06:45:32AM 0 points [-]

Well, orgasmium, for a start.

Comment author: MugaSofer 27 November 2012 05:33:13AM 0 points [-]

must admit I am a tad confused here, but intelligence or whatever seems a good rule of thumb.

I was asking questions to try to better understand where you're coming from. Do you mean the questions were confusing?

No, I mean I am unsure as to what my CEV would answer.

Are you saying that moral worth is directly proportional to intelligence? If so, why do you think this is true?

Because I'll kill a bug to save a chicken, a chicken to save a cat, a cat to save an ape, and an ape to save a human. The part of me responsible for morality clearly has some sort of criteria for moral worth that seems roughly equivalent to intelligence.

But introspection suggests I don't value it as much depending on how conscious they are (probably the same as intelligence.)

Why not? Do you have a good reason, or are you just going off of intuition?

... both?

Have you read "Not for the Sake of Happiness (Alone)"?

Yes, I've read it. I'm not entirely convinced that all values reduce to happiness, but I've never seen any value that can't be reduced to happiness. That's one of the areas in ethics where I'm the most uncertain. In practice, it doesn't come up much because in almost every situation, happiness and preference satisfaction amount to the same thing.

Fair enough. Unfortunately, the area of ethics where I'm the most uncertain is weighting creatures with different intelligence levels.

Thing like discovery and creativity seem like good examples of preferences that don't reduce to happiness IIRC, although it's been a while since I thought everything reduced to happiness so I don't recall very well.

I'm inclined to believe that not all preferences reduce to happiness, but all CEV preferences do reduce to happiness. As I said before, I'm fairly uncertain about this and I don't have much evidence.

Not sure what this means.

Comment author: MTGandP 27 November 2012 05:59:20AM 1 point [-]

Are you saying that moral worth is directly proportional to intelligence? If so, why do you think this is true?

Because I'll kill a bug to save a chicken, a chicken to save a cat, a cat to save an ape, and an ape to save a human. The part of me responsible for morality clearly has some sort of criteria for moral worth that seems roughly equivalent to intelligence.

But why is intelligence important? I don't see its connection to morality. I know it's commonly believed that intelligence is morally relevant, and my best guess as to why is that it conveniently places humans at the top and thus justifies mistreating non-human animals.

If intelligence is morally significant, then it's not really that bad to torture a mentally handicapped person.

I believe this is false: a mentally handicapped person suffers physical pain to the same extent that I do, so his suffering is just as morally significant. The same reasoning applies to many species of non-human animal. What matters is not intelligence but the capacity to experience happiness and suffering.

... both?

So then what is your good reason that's not directly based on intuition?

Thing like discovery and creativity seem like good examples of preferences that don't reduce to happiness IIRC, although it's been a while since I thought everything reduced to happiness so I don't recall very well.

Discovery leads to the invention of new things. In general, new things lead to increased happiness. It also leads to a better understanding of the universe, which allows us to better increase happiness. If the process of discovery brought no pleasure in itself and also didn't make it easier for us to increase happiness, I think it would be useless. The same reasoning applies to creativity.

Not sure what this means.

You mentioned CEV in your previous comment, so I assume you're familiar with it. I mean that I think if you took people's coherent extrapolated volitions, they would exclusively value happiness.

Comment author: MugaSofer 27 November 2012 06:43:35AM 0 points [-]

I'll kill a bug to save a chicken, a chicken to save a cat, a cat to save an ape, and an ape to save a human. The part of me responsible for morality clearly has some sort of criteria for moral worth that seems roughly equivalent to intelligence.

But why is intelligence important? I don't see its connection to morality. I know it's commonly believed that intelligence is morally relevant, and my best guess as to why is that it conveniently places humans at the top and thus justifies mistreating non-human animals.

Well, why is pain important? I suspect empathy is mixed up here somewhere, but honestly, it doesn't feel like it reduces - bugs just are worth less. Besides, where do you draw the line if you lack a sliding scale - I assume you don't care about rocks, or sponges, or germs.

If intelligence is morally significant, then it's not really that bad to torture a mentally handicapped person.

Well ... not as bad as torturing, say, Bob, the Entirely Average Person, no. But it's risky to distinguish between humans like this because it lets in all sorts of nasty biases, so I try not to except in exceptional cases.

I believe this is false: a mentally handicapped person suffers physical pain to the same extent that I do, so his suffering is just as morally significant. The same reasoning applies to many species of non-human animal. What matters is not intelligence but the capacity to experience happiness and suffering.

I know you do. Of course, unless they're really handicapped, most animals are still much lower; and, of course there's the worry that the intelligence is ther and they just can't express it in everyday life (idiot savants and so on.)

So then what is your good reason that's not directly based on intuition?

Well, it's morality, it does ultimately come down to intuition no matter what. I can come up with all sorts of reasons, but remember that they aren't my true rejection - my true rejection is the mental image of killing a man to save some cockroaches.

Discovery leads to the invention of new things. In general, new things lead to increased happiness. It also leads to a better understanding of the universe, which allows us to better increase happiness. If the process of discovery brought no pleasure in itself and also didn't make it easier for us to increase happiness, I think it would be useless. The same reasoning applies to creativity.

And yet, a world without them sounds bleak and lacking in utility.

You mentioned CEV in your previous comment, so I assume you're familiar with it. I mean that I think if you took people's coherent extrapolated volitions, they would exclusively value happiness

Oh, right.

Ah ... not sure what I can say to convince you if NFTSOH(A) didn't.

Comment author: MTGandP 28 November 2012 06:31:26AM 1 point [-]

Well, why is pain important?

It's really abstract and difficult to explain, so I probably won't do a very good job. Peter Singer explains it pretty well in "All Animals Are Equal." Basically, we should give equal consideration to the interests of all beings. Any being capable of suffering has an interest in avoiding suffering. A more intelligent being does not have a greater interest in avoiding suffering [1]; hence, intelligence is not morally relevant.

Besides, where do you draw the line if you lack a sliding scale - I assume you don't care about rocks, or sponges, or germs.

There is a sliding scale. More capacity to feel happiness and suffering = more moral worth. Rocks, sponges, and germs have no capacity to feel happiness and suffering.

And yet, a world without [discovery] sounds bleak and lacking in utility.

Well yeah. That's because discovery tends to increase happiness. But if it didn't, it would be pointless. For example, suppose you are tasked with sifting through a pile of sand to find which one is the whitest. When you finish, you will have discovered something new. But the process is really boring and it doesn't benefit anyone, so what's the point? Discovery is only worthwhile if it increases happiness in some way.

I'm not saying that it's impossible to come up with an example of something that's not reducible to happiness, but I don't think discovery is such a thing.

[1] Unless it is capable of greater suffering, but that's not a trait inherent to intelligence. I think it may be true in some respects that more intelligent beings are capable of greater suffering; but what matters is the capacity to suffer, not the intelligence itself.

Comment author: Jayson_Virissimo 28 November 2012 07:13:10AM *  2 points [-]

There is a sliding scale. More capacity to feel happiness and suffering = more moral worth. Rocks, sponges, and germs have no capacity to feel happiness and suffering.

This sounds like a bad rule and could potentially create a sensitivity arms race. Assuming that people that practice Stoic or Buddhist techniques are successful in diminishing their capacity to suffer, does that mean they are worth less morally than before they started? This would be counter-intuitive, to say the least.

Comment author: MTGandP 29 November 2012 12:05:27AM 1 point [-]

Assuming that people that practice Stoic or Buddhist techniques are successful in diminishing their capacity to suffer, does that mean they are worth less morally than before they started?

It means that inducing some typically-harmful action on a Stoic is less harmful than inducing it on a normal person. For example, suppose you have a Stoic who no longer feels negative reactions to insults. If you insult her, she doesn't mind at all. It would be morally better to insult this person than to insult a typical person.

Let me put it this way: all suffering of equal degree is equally important, and the importance of suffering is proportional to its degree.

A lot of conclusions follow from this principle, including:

  • animal suffering is important;
  • if you have to do something to one of two beings and it will cause greater suffering to being A, then, all else being equal, you should do it to being B.
Comment author: MugaSofer 29 November 2012 09:53:43PM 0 points [-]

Well, why is pain important?

It's really abstract and difficult to explain, so I probably won't do a very good job. Peter Singer explains it pretty well in "All Animals Are Equal." Basically, we should give equal consideration to the interests of all beings. Any being capable of suffering has an interest in avoiding suffering. A more intelligent being does not have a greater interest in avoiding suffering [1]; hence, intelligence is not morally relevant.

No, my point was that your valuing pain is itself a moral intuition. Picture a pebblesorter explaining that this pile is correct, while your pile is, obviously, incorrect.

There is a sliding scale. More capacity to feel happiness and suffering = more moral worth. Rocks, sponges, and germs have no capacity to feel happiness and suffering.

So, say, an emotionless AI? A human with damaged pain receptors? An alien with entirely different neurochemistry analogs?

Well yeah. That's because discovery tends to increase happiness. But if it didn't, it would be pointless. For example, suppose you are tasked with sifting through a pile of sand to find which one is the whitest. When you finish, you will have discovered something new. But the process is really boring and it doesn't benefit anyone, so what's the point? Discovery is only worthwhile if it increases happiness in some way.

No. I'm saying that I value exporation/discovery/whatever even when it serves no purpose, ultimately. Joe may be exploring a randomly-generated landscape, but it's better than sitting in a whitewashed room wireheading nonetheless.

[1] Unless it is capable of greater suffering, but that's not a trait inherent to intelligence. I think it may be true in some respects that more intelligent beings are capable of greater suffering; but what matters is the capacity to suffer, not the intelligence itself.

Can you taboo "suffering" for me?

Comment author: MTGandP 30 November 2012 02:54:45AM 2 points [-]

I've avoided using the word "suffering" or its synonyms in this comment, except in one instance where I believe it is appropriate.

No, my point was that your valuing pain is itself a moral intuition.

Yes, it's an intuition. I can't prove that suffering is important.

So, say, an emotionless AI?

If the AI does not consciously prefer any state to any other state, then it has no moral worth.

A human with damaged pain receptors?

Such a human could still experience emotions, so ey would still have moral worth.

An alien with entirely different neurochemistry analogs?

Difficult to say. If it can experience states about which it has an interest in promoting or avoiding, then it has moral worth.

No. I'm saying that I value exporation/discovery/whatever even when it serves no purpose, ultimately. Joe may be exploring a randomly-generated landscape, but it's better than sitting in a whitewashed room wireheading nonetheless.

Okay. I don't really get why, but I can't dispute that you hold that value. This is why preference utilitarianism can be nice.

Comment author: MugaSofer 30 November 2012 09:21:30AM *  1 point [-]

... oh.

You were defining pain/suffering/whatever as generic disutility? That's much more reasonable.

... so, is a hive of bees one mind of many or sort of both at once? Does evolution get a vote, here? If you aren't discounting optimizers that lack consciousness you're gonna get some damn strange results with this.