Category Archives: 1

How far can AI jump?

Posted on October 16, 2009 | 47 comments

I went to the Singularity Summit recently, organized by the Singularity Institute for Artificial Intelligence (SIAI). SIAI’s main interest is in the prospect of a superintelligence quickly emerging and destroying everything we care about in the reachable universe. This concern has two components. One is that any AI above ‘human level’ will improve its intelligence further until it takes over the world from all other entities. The other is that when the intelligence that takes off is created it will accidentally have the wrong values, and because it is smart and thus very good at bringing about what it wants, it will destroy all that humans value. I disagree that either part is likely. Here I’ll summarize why I find the first part implausible, and there I discuss the second part.

The reason that an AI – or a group of them – is a contender for gaining existentially risky amounts of power is that it could trigger an intelligence explosion which happens so fast that everyone else is left behind. An intelligence explosion is a positive feedback where more intelligent creatures are better at improving their intelligence further.

Such a feedback seems likely. Even now as we gain more concepts and tools that allow us to think well we use them to make more such understanding. AIs fiddling with their architecture don’t seem fundamentally different. But feedback effects are easy to come by. The question is how big this feedback effect will become. Will it be big enough for one machine to permanently overtake the rest of the world economy in accumulating capability?

In order to grow more powerful than everyone else you need to get significantly ahead at some point. You can imagine this could happen either by having one big jump in progress or by having slightly more growth over a long period of time. Having slightly more growth over a long period is staggeringly unlikely to happen by chance, so it needs to share some cause too. Anything that will give you higher growth for long enough to take over the world is a pretty neat innovation, and for you to take over the world everyone else has to not have anything close. So again, this is a big jump in progress. So for AI to help a small group take over the world, it needs to be a big jump.

Notice that no jumps have been big enough before in human invention. Some species, such as humans, have mostly taken over the worlds of other species. The seeming reason for this is that there was virtually no sharing of the relevant information between species. In human society there is a lot of information sharing. This makes it hard for anyone to get far ahead of everyone else. While you can see there are barriers to insights passing between groups, such as incompatible approaches to a kind of technology by different people working on it, these have not so far caused anything like a gap allowing permanent separation of one group.

Another barrier to a big enough jump is that much human progress comes from the extra use of ideas that sharing information brings. You can imagine that if someone predicted writing they might think ‘whoever creates this will be able to have a superhuman memory and accumulate all the knowledge in the world and use it to make more knowledge until they are so knowledgeable they take over everything.’ If somebody created writing and kept it to themselves they would not accumulate nearly as much recorded knowledge as another person who shared a writing system. The same goes for most technology. At the extreme, if nobody shared information, each person would start out with less knowledge than a cave man, and would presumably end up with about that much still. Nothing invented would be improved on. Systems which are used tend to be improved on more. This means if a group hides their innovations and tries to use them alone to create more innovation, the project will probably not grow as fast as the rest of the economy together. Even if they still listen to what’s going on outside, and just keep their own innovations secret, a lot of improvement in technologies like software comes from use. Forgoing information sharing to protect your advantage will tend to slow down your growth.

Those were some barriers to an AI project causing a big enough jump. Are the reasons for it good enough to make up for them?

The main argument for an AI jump seems to be that human level AI is a powerful and amazing innovation that will cause a high growth rate. But this means it is a leap from what we have currently, not that it is especially likely to be arrived at in one leap. If we invented it tomorrow it would be a jump, but that’s just evidence that we won’t invent it tomorrow. You might argue here that however gradually it arrives, the AI will be around human level one day, and then the next it will suddenly be a superpower. There’s a jump from the growth after human level AI is reached, not before. But if it is arrived at incrementally then others are likely to be close in developing similar technology, unless it is a secret military project or something. Also an AI which recursively improves itself forever will probably be preceded by AIs which self improve to a lesser extent, so the field will be moving fast already. Why would the first try at an AI which can improve itself have infinite success? It’s true that if it were powerful enough it wouldn’t matter if others were close behind or if it took the first group a few goes to make it work. For instance if it only took a few days to become as productive as the rest of the world added together, the AI could probably prevent other research if it wanted. However I haven’t heard any good evidence it’s likely to happen that fast.

Another argument made for an AI project causing a big jump is that intelligence might be the sort of thing for which there is a single principle. Until you discover it you have nothing, and afterwards you can build the smartest thing ever in an afternoon and can just extend it indefinitely. Why would intelligence have such a principle? I haven’t heard any good reason. That we can imagine a simple, all powerful principle of controlling everything in the world isn’t evidence for it existing.

I agree human level AI will be a darn useful achievement and will probably change things a lot, but I’m not convinced that one AI or one group using it will take over the world, because there is no reason it will be a never before seen size jump from technology available before it.

47 Comments

Posted in 1

Tagged , ,

Dominant characters on the left

Posted on October 1, 2009 | 5 comments

From Psyblog:

Research finds that people or objects moving from left to right are perceived as having greater power (Maass et al., 2007):

  • Soccer goals are rated as stronger, faster, even more beautiful when the movement of the scorer is from left to right, rather than right to left.
  • Film violence seems more aggressive, more painful and more shocking when the punch is delivered from left to right, compared with right to left.
  • Cars in an advert are rated as stronger and faster when they are moving from left to right, rather than right to left (take note advertising executives!).

Perhaps it’s no coincidence that athletes, cars and horses are all usually shown on TV reaching the finishing line from left to right.

According to other studies mentioned by Maass et al, in Western societies most people also tend to preferentially imagine events evolving from left to right, picture the situations in subject-verb-object sentences with the subject on the left of the object, look at new places rather than old ones more when stimuli show up in a left to right order, memorize the final positions of objects further along the implied path more when they are moving left to right, imagine number lines and time increasing from left to right, and scan their eyes over art in a left to right trajectory.

Why is this?

It seems likely that this left to right bias has its roots in language… people who speak languages written from right to left like Arabic or Urdu … display the same bias, but in the opposite direction.

So Maass and the others guessed that characters perceived as more active would also tend to be depicted on the left of more passive characters in pictures. Their research agreed:

We propose that spatial imagery is systematically linked to stereotypic beliefs, such that more agentic groups are envisaged to the left of less agentic groups. This spatial agency bias was tested in three studies. In Study 1, a content analysis of over 200 images of male–female pairs (including artwork, photographs, and cartoons) showed that males were over-proportionally presented to the left of females, but only for couples in which the male was perceived as more agentic. Study 2 (N = 40) showed that people tend to draw males to the left of females, but only if they hold stereotypic beliefs that associate males with greater agency. Study 3 (N = 61) investigated whether scanning habits due to writing direction are responsible for the spatial agency bias. We found a tendency for Italian-speakers to position agentic groups (men and young people) to the left of less agentic groups (females and old people), but a reversal in Arabic-speakers who tended to position the more agentic groups to the right. Together, our results suggest a subtle spatial bias in the representation of social groups that seems to be linked to culturally determined writing/reading habits.

Adam appeared on the left in 62% of paintings considered, far less than Gomez Addams is portrayed on the left of Mortissa (82%). (Picture: Peter Paul Rubens)

Adam appeared on the left in 62% of paintings considered, far less than Gomez Addams is portrayed on the left of Mortissa (82%). (Picture: Peter Paul Rubens)

Note that the first study only looked at four couples; Adam and Eve, Gomez and Mortissa Addams, Fred and Wilma Flinstone, and Marge and Homer Simpson. The last three were compared to surveyed opinions on the couple’s relative activeness, dominance and communion, and the Flinstone and Simpson couples were found to be about equal. An earlier study also found that Gabriel was portrayed to the left of Mary 97% of the time.

Most languages also mention the (active) subject before the object, which means the active entity is on the left when written in a left-to-right language. Grammar predates writing, so if this ordering of nouns is relevant, as the researchers suggest, it seems it combines with the direction of writing to cause the left to right bias. It would be interesting to see whether natives to the few languages who put the object before the subject have this bias  in the other direction. It would also be interesting to see whether the layout of sentences more commonly influences our perceptions of the content, or whether the effect is so weak as to only have influence over years of parsing the same patterns.

5 Comments

Posted in 1

Tagged , , , ,

‘Clearly’ covers murky thought

Posted on September 22, 2009 | 17 comments

Why do we point out that statements we are making are obvious? If a statement is actually obvious, there should rarely reason to point the statement out, let alone that it is obvious. It’s obviousness should be obvious. It seems that a person often emphasizes that a statement is obvious when they would prefer not be required to defend it. Sometimes this is just because it is obvious once you know their field but a lot of effort to explain to someone who doesn’t, but often it’s just that the explanation is not obvious to them.

But saying ‘obviously’ is too obvious. A better word is ‘clearly’. ‘Clearly’ sounds transparent and innocent. In reality it is a more subtle version of ‘obviously’.

I have noticed this technique used well in published philosophy from time to time. If getting to your conclusion is going to require assuming your conclusion is true, ‘clearly’ suggests to the reader that they not think over that step too closely.

For instance Michael Huemer in Ethical Intuitionism, while arguing that moral subjectivism is wrong, for the purpose of demonstrating that ethical intuitionism is right:

Traditionally, cultural relativists have been charged with endorsing such statements as,

If society were to approve of eating children, then eating children would be good.

which is clearly false.

Notice that ‘false’ seemingly means that it is false according to his intuition; the thing which he is trying to argue for the reliability of. If he just said ‘which is false’, the reader may wonder where, in a book on establishing a basis for ethical truth, this source of falsity may have popped from. ‘Clearly’ says that they needn’t worry about it.

17 Comments

Posted in 1

Tagged , ,

Why do animal lovers want animals to feel pain?

Posted on September 21, 2009 | 37 comments
Behind the vail of (lots of) ignorance, would you rather squished chickens be painless?

Behind the vail of (lots of) ignorance, would you rather squished chickens be painless?

We may soon be able to make pain-free animals, according to New Scientist. The study they reported on finds that people not enthused by creating such creatures for scientific research, which is interesting. Robin Hanson guessed prior to seeing the article that this was because endorsing pain free animals would require thinking that farmed animals now were in more pain than wild animals, which people don’t think. However it turns out that vegetarians and animal welfare advocates were much more opposed to the idea than others in the study, so another explanation is needed.

Robert Wiblin suggested to me that vegetarians are mostly in favor of animals not being used, as well as not being hurt, so they don’t want to support pain-free use, as that is supporting use. He made this comparison:

Currently children are being sexually abused. The technology now exists to put them under anaesthetic so that they don’t experience the immediate pain of sexual abuse. Should we put children under anaesthetic to sexually abuse them?

A glance at the comments on other sites reporting the possibility of painless meat suggests vegetarians cite this along with a lot of different reasons for disapproval. And sure enough it seems mainly meat eaters who say eliminating pain would make them feel better about eating meat. The reasons vegetarians (and others) give for not liking the idea, or for not being more interested in pain-free meat, include:

  • The animals would harm themselves without knowing
  • Eating animals is bad for environmental or health reasons
  • Killing is always wrong
  • Animals have complex social lives and are sad when their family are killed, regardless of pain
  • Animals are living things [?!]
  • There are other forms of unpleasantness, such as psychological torture
  • How can we tell they don’t feel pain?
  • We will treat them worse if we think they can’t feel it, and we might be wrong
  • There are better solutions, such as not eating meat
  • It’s weird, freaky, disrespectful
  • It’s selfish and unnecessary for humans to do this to animals

Many reasonable reasons. The fascinating thing though is that vegetarians seem to consistently oppose the idea, yet not share reasons. Three (not mutually exclusive) explanations:

  1. Vegetarians care more about animals in general, so care about lots of related concerns.
  2. Once you have an opinion, you collect a multitude of reasons to have it. When I was a vegetarian I thought meat eating was bad for the environment, bad for people who need food, bad for me, maybe even bad for animals. This means when a group of people lose one reason to hold a shared belief they all have other reasons to put forward, but not necessarily the same ones.
  3. There’s some single reason vegetarians are especially motivated to oppose pain-free meat, so they each look for a reason to oppose it, and come across different ones, as there are many.

I’m interested by 3 because the situation reminds me of a pattern in similar cases I have noticed before. It goes like this. Some people make personal sacrifices, supposedly toward solving problems that don’t threaten them personally. They sort recycling, buy free range eggs, buy fair trade, campaign for wealth redistribution etc. Their actions are seen as virtuous. They see those who don’t join them as uncaring and immoral. A more efficient solution to the problem is suggested. It does not require personal sacrifice. People who have not previously sacrificed support it. Those who have previously sacrificed object on grounds that it is an excuse for people to get out of making the sacrifice.

The supposed instrumental action, as the visible sign of caring, has become virtuous in its own right. Solving the problem effectively is an attack on the moral people – an attempt to undermine their dream of a future where everybody longs to be very informed on the social and environmental effects of their consumption choices or to sort their recycling really well. Some examples of this sentiment:

  • A downside to recreating extinct species with cloning is that it will let people bother even less about stopping extinctions.
  • A recycling system where items are automatically and efficiently sorted at the plant rather than individually in homes would be worse because then people would be ignorant about the effort it takes to recycle.
  • Modern food systems lamentably make people lazy and ignorant of where their food comes from.
  • Making cars efficient just lets people be lazy and drive them more, rather than using real solutions like bikes.
  • The internet’s ready availability and general knowledge allows people to be ignorant and not bother learning facts.

In these cases, having solved a problem a better way should mean that efforts to solve it via personal sacrifice can be lessened. This would be a good thing if we wanted to solve the problem, and didn’t want to sacrifice. We would rejoice at progress allowing ever more ignorance and laziness on a given issue. But often we instead regret the end of an opportunity to show compassion and commitment. Especially when we were the compassionate, committed ones.

Is vegetarian opposition to preventing animal pain an example of this kind of motivation? Vegetarianism is a big personal effort, a moral issue, a cause of feelings of moral superiority, and a feature of identity which binds people together. It looks like other issues where people readily claim fear of an end to virtuous efforts.  How should we distinguish between this and the other explanations?

37 Comments

Posted in 1

Tagged , , , ,

Charitable explanation

Posted on September 8, 2009 | 15 comments

Is anyone really altruistic? The usual cynical explanations for seemingly altruistic behavior are that it makes one feel good, it makes one look good, and it brings other rewards later. These factors are usually present, but how much do they contribute to motivation?

One way to tell if it’s all about altruism is to invite charity that explicitly won’t benefit anyone. Curious economists asked their guinea pigs for donations to a variety of causes, warning them:

“The amount contributed by the proctor to your selected charity WILL be reduced by however much you pass to your selected charity. Your selected charity will receive neither more nor less than $10.”

Many participants chipped in nonetheless:

We find that participants, on average, donated 20% of their endowments and that approximately 57% of the participants made a donation.

This is compared to giving an average of 30-49% in experiments where donating benefited the cause, but it is of course possible that knowing you are helping offers more of a warm glow. It looks like at least half of giving isn’t altruistic at all, unless the participants were interested in the wellbeing of the experimenters’ funds.

The opportunity to be observed by others also influences how much we donate, and we are duly rewarded with reputation:

Here we demonstrate that more subjects were willing to give assistance to unfamiliar people in need if they could make their charity offers in the presence of their group mates than in a situation where the offers remained concealed from others. In return, those who were willing to participate in a particular charitable activity received significantly higher scores than others on scales measuring sympathy and trustworthiness.

This doesn’t tell us whether real altruism exists though. Maybe there are just a few truly altruistic deeds out there? What would a credibly altruistic act look like?

Fortunately for cute children desirous of socially admirable help, much charity is not driven by altruism (picture: Laura Lartigue)

Fortunately for cute children desirous of socially admirable help, much charity is not driven by altruism (picture: Laura Lartigue)

If an act made the doer feel bad, look bad to others, and endure material cost, while helping someone else, we would probably be satisfied that it was altruistic. For instance if a person killed their much loved grandmother to steal her money to donate to a charity they believed would increase the birth rate somewhere far away, at much risk to themselves, it would seem to escape the usual criticisms. And there is no way you would want to be friends with them.

So why would anyone tell you if they had good evidence they had been altruistic? The more credible evidence should look particularly bad. And if they were keen to tell you about it anyway, you would have to wonder whether it was for show after all. This makes it hard for an altruist to credibly inform anyone that they were altruistic. On the other hand the non-altruistic should be looking for any excuse to publicize their good deeds. This means the good deeds you hear about should be very biased toward the non-altruistic. Even if altruism were all over the place it should be hard to find. But it’s not, is it?

15 Comments

Posted in 1

Tagged , ,