Less Wrong/2009 Articles/Summaries
It may be better to create a world that operates by better rules, that you can understand, so that you can optimize your own future, than to create a world that includes some sort of deity that can be prayed to. The human reluctance to have their future controlled by an outside source is a nontrivial part of morality.
Fun Theory is important for replying to critics of human progress; for inspiring people to keep working on human progress; for refuting religious arguments that the world could possibly have been benevolently designed; for showing that religious Heavens show the signature of the same human biases that torpedo other attempts at Utopia; and for appreciating the great complexity of our values and of a life worth living, which requires a correspondingly strong effort of AI design to create AIs that can play good roles in a good future.
Each part of the human brain is optimized for behaving correctly, assuming that the rest of the brain is operating exactly as expected. Change one part, and the rest of your brain may not work as well. Increasing a human's intelligence is not a trivial problem.
Creating new emotions seems like a desirable aspect of many parts of Fun Theory, but this is not to be trivially postulated. It's the sort of thing best done with superintelligent help, and slowly and conservatively even then. We can illustrate these difficulties by trying to translate the short English phrase "change sex" into a cognitive transformation of extraordinary complexity and many hidden subproblems.
Since the events in video games have no actual long-term consequences, playing a video game is not likely to be nearly as emotionally involving as much less dramatic events in real life. The supposed Utopia of playing lots of cool video games forever, is life as a series of disconnected episodes with no lasting consequences. Our current emotions are bound to activities that were subgoals of reproduction in the ancestral environment - but we now pursue these activities as independent goals regardless of whether they lead to reproduction.
Stories and lives are optimized according to rather different criteria. Advice on how to write fiction will tell you that "stories are about people's pain" and "every scene must end in disaster". I once assumed that it was not possible to write any story about a successful Singularity because the inhabitants would not be in any pain; but something about the final conclusion that the post-Singularity world would contain no stories worth telling seemed alarming. Stories in which nothing ever goes wrong, are painful to read; would a life of endless success have the same painful quality? If so, should we simply eliminate that revulsion via neural rewiring? Pleasure probably does retain its meaning in the absence of pain to contrast it; they are different neural systems. The present world has an imbalance between pain and pleasure; it is much easier to produce severe pain than correspondingly intense pleasure. One path would be to address the imbalance and create a world with more pleasures, and free of the more grindingly destructive and pointless sorts of pain. Another approach would be to eliminate pain entirely. I feel like I prefer the former approach, but I don't know if it can last in the long run.
Humans seem to be on a hedonic treadmill; over time, we adjust to any improvements in our environment - after a month, the new sports car no longer seems quite as wonderful. This aspect of our evolved psychology is not surprising: it is a rare organism in a rare environment whose optimal reproductive strategy is to rest with a smile on its face, feeling happy with what it already has. To entirely delete the hedonic treadmill seems perilously close to tampering with Boredom itself. Is there enough fun in the universe for a transhuman to jog off the treadmill - improve their life continuously, leaping to ever-higher hedonic levels before adjusting to the previous one? Can ever-higher levels of pleasure be created by the simple increase of ever-larger floating-point numbers in a digital pleasure center, or would that fail to have the full subjective quality of happiness? If we continue to bind our pleasures to novel challenges, can we find higher levels of pleasure fast enough, without cheating? The rate at which value can increase as more bits are added, and the rate at which value must increase for eudaimonia, together determine the lifespan of a mind. If minds must use exponentially more resources over time in order to lead a eudaimonic existence, their subjective lifespan is measured in mere millennia even if they can draw on galaxy-sized resources.
If a citizen of the Past were dropped into the Present world, they would be pleasantly surprised along at least some dimensions; they would also be horrified, disgusted, and frightened. This is not because our world has gone wrong, but because it has gone right. A true Future gone right would, realistically, be shocking to us along at least some dimensions. This may help explain why most literary Utopias fail; as George Orwell observed, "they are chiefly concerned with avoiding fuss". Heavens are meant to sound like good news; political utopias are meant to show how neatly their underlying ideas work. Utopia is reassuring, unsurprising, and dull. Eutopia would be scary. (Of course the vast majority of scary things are not Eutopian, just entropic.) Try to imagine a genuinely better world in which you would be out of place - not a world that would make you smugly satisfied at how well all your current ideas had worked. This proved to be a very important exercise when I tried it; it made me realize that all my old proposals had been optimized to sound safe and reassuring.
Utopia and Dystopia both confirm the moral sensibilities you started with; whether the world is a libertarian utopia of government non-interference, or a hellish dystopia of government intrusion and regulation, either way you get to say "Guess I was right all along." To break out of this mold, write down the Utopia, and the Dystopia, and then try to write down the Weirdtopia - an arguably-better world that zogs instead of zigging or zagging. (Judging from the comments, this exercise seems to have mostly failed.)
A pleasant surprise probably has a greater hedonic impact than being told about the same positive event long in advance - hearing about the positive event is good news in the moment of first hearing, but you don't have the gift actually in hand. Then you have to wait, perhaps for a long time, possibly comparing the expected pleasure of the future to the lesser pleasure of the present. This argues that if you have a choice between a world in which the same pleasant events occur, but in the first world you are told about them long in advance, and in the second world they are kept secret until they occur, you would prefer to live in the second world. The importance of hope is widely appreciated - people who do not expect their lives to improve in the future are less likely to be happy in the present - but the importance of vague hope may be understated.
Vagueness usually has a poor name in rationality, but the Future is something about which, in fact, we do not possess strong reliable specific information. Vague (but justified!) hopes may also be hedonically better. But a more important caution for today's world is that highly specific pleasant scenarios can exert a dangerous power over human minds - suck out our emotional energy, make us forget what we don't know, and cause our mere actual lives to pale by comparison. (This post is not about Fun Theory proper, but it contains an important warning about how not to use Fun Theory.)
How should a rationalist use their near and far modes of thinking? And how should knowing about near vs far modes influence how we present the things we believe to other people.
"Boredom" is an immensely subtle and important aspect of human values, nowhere near as straightforward as it sounds to a human. We don't want to get bored with breathing or with thinking. We do want to get bored with playing the same level of the same video game over and over. We don't want changing the shade of the pixels in the game to make it stop counting as "the same game". We want a steady stream of novelty, rather than spending most of our time playing the best video game level so far discovered (over and over) and occasionally trying out a different video game level as a new candidate for "best". These considerations would not arise in most utility functions in expected utility maximizers.
Mirror neurons are neurons that fire both when performing an action oneself, and watching someone else perform the same action - for example, a neuron that fires when you raise your hand or watch someone else raise theirs. We predictively model other minds by putting ourselves in their shoes, which is empathy. But some of our desire to help relatives and friends, or be concerned with the feelings of allies, is expressed as sympathy, feeling what (we believe) they feel. Like "boredom", the human form of sympathy would not be expected to arise in an arbitrary expected-utility-maximizing AI. Most such agents would regard any agents in its environment as a special case of complex systems to be modeled or optimized; it would not feel what they feel.
Our sympathy with other minds makes our interpersonal relationships one of the most complex aspects of human existence. Romance, in particular, is more complicated than being nice to friends and kin, negotiating with allies, or outsmarting enemies - it contains aspects of all three. Replacing human romance with anything simpler or easier would decrease the peak complexity of the human species - a major step in the wrong direction, it seems to me. This is my problem with proposals to give people perfect, nonsentient sexual/romantic partners, which I usually refer to as "catgirls" ("catboys"). The human species does have a statistical sex problem: evolution has not optimized the average man to make the average woman happy or vice versa. But there are less sad ways to solve this problem than both genders giving up on each other and retreating to catgirls/catboys.
A fictional short story illustrating some of the ideas in Interpersonal Entanglement above. (Many commenters seemed to like this story, and some said that the ideas were easier to understand in this form.)
Having a Purpose in Life consistently shows up as something that increases stated well-being. Of course, the problem with trying to pick out "a Purpose in Life" in order to make yourself happier, is that this doesn't take you outside yourself; it's still all about you. To find purpose, you need to turn your eyes outward to look at the world and find things there that you care about - rather than obsessing about the wonderful spiritual benefits you're getting from helping others. In today's world, most of the highest-priority legitimate Causes consist of large groups of people in extreme jeopardy: Aging threatens the old, starvation threatens the poor, extinction risks threaten humanity as a whole. If the future goes right, many and perhaps all such problems will be solved - depleting the stream of victims to be helped. Will the future therefore consist of self-obsessed individuals, with nothing to take them outside themselves? I suggest, though, that even if there were no large groups of people in extreme jeopardy, we would still, looking around, find things outside ourselves that we cared about - friends, family; truth, freedom... Nonetheless, if the Future goes sufficiently well, there will come a time when you could search the whole of civilization, and never find a single person so much in need of help, as dozens you now pass on the street. If you do want to save someone from death, or help a great many people, then act now; your opportunity may not last, one way or another.
Describes some of the many complex considerations that determine what sort of happiness we most prefer to have - given that many of us would decline to just have an electrode planted in our pleasure centers.
describes some of the many complex considerations that determine what sort of happiness we most prefer to have - given that many of us would decline to just have an electrode planted in our pleasure centers.
You should try hard and often to test your rationality, but how can you do that?
Deceiving yourself is harder than it seems. What looks like a successively adopted false belief may actually be just a belief in false belief.
Behind every particular failure of social rationality is a larger and more general failure of social rationality; even if all religious content were deleted tomorrow from all human minds, the larger failures that permit religion would still be present. Religion may serve the function of an asphyxiated canary in a coal mine - getting rid of the canary doesn't get rid of the gas. Even a complete social victory for atheism would only be the beginning of the real work of rationalists. What could you teach people without ever explicitly mentioning religion, that would raise their general epistemic waterline to the point that religion went underwater?
Why do people seem to care more about systematic methods of punching than systematic methods of thinking?
The art of human rationality may have not been much developed because its practitioners lack a sense that vastly more is possible. The level of expertise that most rationalists strive to develop is not on a par with the skills of a professional mathematician - more like that of a strong casual amateur. Self-proclaimed "rationalists" don't seem to get huge amounts of personal mileage out of their craft, and no one sees a problem with this. Yet rationalists get less systematic training in a less systematic context than a first-dan black belt gets in hitting people.
The classic failure modes of martial arts schools transfer over as direct warnings to would-be rationalists.
An essay by Gillian Russell on "Epistemic Viciousness in the Martial Arts" generalizes amazingly to possible and actual problems with building a community around rationality. Most notably the extreme dangers associated with "data poverty" - the difficulty of testing the skills in the real world. But also such factors as the sacredness of the dojo, the investment in teachings long-practiced, the difficulty of book learning that leads into the need to trust a teacher, deference to historical masters, and above all, living in data poverty while continuing to act as if the luxury of trust is possible.
The branching schools of "psychotherapy", another domain in which experimental verification was weak (nonexistent, actually), show that an aspiring craft lives or dies by the degree to which it can be tested in the real world. In the absence of that testing, one becomes prestigious by inventing yet another school and having students, rather than excelling at any visible performance criterion. The field of hedonic psychology (happiness studies) began, to some extent, with the realization that you could measure happiness - that there was a family of measures that by golly did validate well against each other. The act of creating a new measurement creates new science; if it's a good measurement, you get good science.
Reputational, experimental, and organizational: the different strengths of evidence needed to (a) prevent a school from degenerating, (b) systematically test particular techniques, or (c) credential individuals.
How far the craft of rationality can be taken, depends largely on what methods can be invented for verifying it. Tests seem usefully stratifiable into reputational, experimental, and organizational. A "reputational" test is some real-world problem that tests the ability of a teacher or a school (like running a hedge fund, say) - "keeping it real", but without being able to break down exactly what was responsible for success. An "experimental" test is one that can be run on each of a hundred students (such as a well-validated survey). An "organizational" test is one that can be used to preserve the integrity of organizations by validating individuals or small groups, even in the face of strong incentives to game the test. The strength of solution invented at each level will determine how far the craft of rationality can go in the real world.
Brainstorming verification tests, asking along what dimensions you think you've improved due to "rationality".
The atheist/libertarian/technophile/sf-fan/early-adopter/programmer/etc crowd, aka "the nonconformist cluster", seems to be stunningly bad at coordinating group projects. There are a number of reasons for this, but one of them is that people are as reluctant to speak agreement out loud, as they are eager to voice disagreements - the exact opposite of the situation that obtains in more cohesive and powerful communities. This is not rational either! It is dangerous to be half a rationalist (in general), and this also applies to teaching only disagreement but not agreement, or only lonely defiance but not coordination. The pseudo-rationalist taboo against expressing strong feelings probably doesn't help either.
One of the likely characteristics of someone who sets out to be a "rationalist" is a lower-than-usual tolerance for flawed thinking. This makes it very important to tolerate other people's tolerance - to avoid rejecting them because they tolerate people you wouldn't - since otherwise we must all have exactly the same standards of tolerance in order to work together, which is unlikely. Even if someone has a nice word to say about complete lunatics and crackpots - so long as they don't literally believe the same ideas themselves - try to be nice to them? Intolerance of tolerance corresponds to punishment of non-punishers, a very dangerous game-theoretic idiom that can lock completely arbitrary systems in place even when they benefit no one at all.
Paul Graham gets exactly the same accusations about "cults" and "echo chambers" and "coteries" that I do, in exactly the same tone - e.g. comparing the long hours worked by Y Combinator startup founders to the sleep-deprivation tactic used in cults, or claiming that founders were asked to move to the Bay Area startup hub as a cult tactic of separation from friends and family. This is bizarre, considering our relative surface risk factors. It just seems to be a failure mode of the nonconformist community in general. By far the most cultish-looking behavior on Hacker News is people trying to show off how willing they are to disagree with Paul Graham, which, I can personally testify, feels really bizarre when you're the target. Admiring someone shouldn't be so scary - I don't hold back so much when praising e.g. Douglas Hofstadter; in this world there are people who have pulled off awesome feats and it is okay to admire them highly.
Seven thoughts: I can list more than one thing that is awesome; when I think of "Douglas Hofstadter" I am really thinking of his all-time greatest work; the greatest work is not the person; when we imagine other people we are imagining their output, so the real Douglas Hofstadter is the source of "Douglas Hofstadter"; I most strongly get the sensation of awesomeness when I see someone outdoing me overwhelmingly, at some task I've actually tried; we tend to admire unique detailed awesome things and overlook common nondetailed awesome things; religion and its bastard child "spirituality" tends to make us overlook human awesomeness.
The game-theoretical puzzle of the Ultimatum game has its reflection in a real-world dilemma: How much do you demand that an existing group adjust toward you, before you will adjust toward it? Our hunter-gatherer instincts will be tuned to groups of 40 with very minimal administrative demands and equal participation, meaning that we underestimate the inertia of larger and more specialized groups and demand too much before joining them. In other groups this resistance can be overcome by affective death spirals and conformity, but rationalists think themselves too good for this - with the result that people in the nonconformist cluster often set their joining prices way way way too high, like an 50-way split with each player demanding 20% of the money. Nonconformists need to move in the direction of joining groups more easily, even in the face of annoyances and apparent unresponsiveness. If an issue isn't worth personally fixing by however much effort it takes, it's not worth a refusal to contribute.
Anyone with a simple and obvious charitable project - responding with food and shelter to a tidal wave in Thailand, say - would be better off by far pleading with the Pope to mobilize the Catholics, rather than with Richard Dawkins to mobilize the atheists. For so long as this is true, any increase in atheism at the expense of Catholicism will be something of a hollow victory, regardless of all other benefits. Can no rationalist match the motivation that comes from the irrational fear of Hell? Or does the real story have more to do with the motivating power of physically meeting others who share your cause, and group norms of participating?
Churches serve a role of providing community - but they aren't explicitly optimized for this, because their nominal role is different. If we desire community without church, can we go one better in the course of deleting religion? There's a great deal of work to be done in the world; rationalist communities might potentially organize themselves around good causes, while explicitly optimizing for community.
Many causes benefit particularly from the spread of rationality - because it takes a little more rationality than usual to see their case, as a supporter, or even just a supportive bystander. Not just the obvious causes like atheism, but things like marijuana legalization. In the case of my own work this effect was strong enough that after years of bogging down I threw up my hands and explicitly recursed on creating rationalists. If such causes can come to terms with not individually capturing all the rationalists they create, then they can mutually benefit from mutual effort on creating rationalists. This cooperation may require learning to shut up about disagreements between such causes, and not fight over priorities, except in specialized venues clearly marked.
Requesting suggestions for an actual survey to be run.
When you consider that our grouping instincts are optimized for 50-person hunter-gatherer bands where everyone knows everyone else, it begins to seem miraculous that modern-day large institutions survive at all. And in fact, the vast majority of large modern-day institutions simply fail to exist in the first place. This is why funding of Science is largely through money thrown at Science rather than donations from individuals - research isn't a good emotional fit for the rare problems that individuals can manage to coordinate on. In fact very few things are, which is why e.g. 200 million adult Americans have such tremendous trouble supervising the 535 members of Congress. Modern humanity manages to put forth very little in the way of coordinated individual effort to serve our collective individual interests.
Omohundro's resource balance principle implies that the inside of any approximately rational system has a common currency of expected utilons. In our world, this common currency is called "money" and it is the unit of how much society cares about something - a brutal yet obvious point. Many people, seeing a good cause, would prefer to help it by donating a few volunteer hours. But this avoids the tremendous gains of comparative advantage, professional specialization, and economies of scale - the reason we're not still in caves, the only way anything ever gets done in this world, the tools grownups use when anyone really cares. Donating hours worked within a professional specialty and paying-customer priority, whether directly, or by donating the money earned to hire other professional specialists, is far more effective than volunteering unskilled hours.
Wealthy philanthropists typically make the mistake of trying to purchase warm fuzzy feelings, status among friends, and actual utilitarian gains, simultaneously; this results in vague pushes along all three dimensions and a mediocre final result. It should be far more effective to spend some money/effort on buying altruistic fuzzies at maximum optimized efficiency (e.g. by helping people in person and seeing the results in person), buying status at maximum efficiency (e.g. by donating to something sexy that you can brag about, regardless of effectiveness), and spending most of your money on expected utilons (chosen through sheer cold-blooded shut-up-and-multiply calculation, without worrying about status or fuzzies).
Trying to breed e.g. egg-laying chickens by individual selection can produce odd side effects on the farm level, since a more dominant hen can produce more egg mass at the expense of other hens. Group selection is nearly impossible in Nature, but easy to impose in the laboratory, and group-selecting hens produced substantial increases in efficiency. Though most of my essays are about individual rationality - and indeed, Traditional Rationality also praises the lone heretic more than evil Authority - the real effectiveness of "rationalists" may end up determined by their performance in groups.
The optimality theorems for probability theory and decision theory, are for perfect probability theory and decision theory. There is no theorem that incremental changes toward the ideal, starting from a flawed initial form, must yield incremental progress at each step along the way. Since perfection is unattainable, why dare to try for improvement? But my limited experience with specialized applications suggests that given enough progress, one can achieve huge improvements over baseline - it just takes a lot of progress to get there.
Many communities feed emotional needs by offering their members someone or something to blame for failure - say, those looters who don't approve of your excellence. You can easily imagine some group of "rationalists" congratulating themselves on how reasonable they were, while blaming the surrounding unreasonable society for keeping them down. But this is not how real rationality works - there's no assumption that other agents are rational. We all face unfair tests (and yes, they are unfair to different degrees for different people); and how well you do with your unfair tests, is the test of your existence. Rationality is there to help you win anyway, not to provide a self-handicapping excuse for losing. There are no first-person extenuating circumstances. There is absolutely no point in going down the road of mutual bitterness and consolation, about anything, ever.
This post was not well-received, but the point was to suggest that a student must at some point leave the dojo and test their skills in the real world. The aspiration of an excellent student should not consist primarily of founding their own dojo and having their own students.
The term "playing to win" comes from Sirlin's book and can be described as using every means necessary to win as long as those means are legal within the structure of the game being played.
Aspiring rationalists often vastly overestimate their own ability to optimize other people's lives. They read nineteen webpages offering productivity advice that doesn't work for them... and then encounter the twentieth page, or invent a new method themselves, and wow, it really works - they've discovered the true method. Actually, they've just discovered the one method in twenty that works for them, and their confident advice is no better than randomly selecting one of the twenty blog posts. Other-Optimizing is exceptionally dangerous when you have power over the other person - for then you'll just believe that they aren't trying hard enough.
The Shangri-La diet works amazingly well for some people, but completely fails for others, for no known reason. Since the diet has a metabolic rationale and is not supposed to require willpower, its failure in my and other cases is unambigiously mysterious. If it required a component of willpower, then I and others might be tempted to blame myself for not having willpower. The art of combating akrasia (willpower failure) has the same sort of mysteries and is in the same primitive state; we don't know the deeper rule that explains why a trick works for one person but not another.
The causes of bystander apathy are even worse on the Internet. There may be an opportunity here for a startup to deliberately try to avert bystander apathy in online group coordination.
Suppose that a country of rationalists is attacked by a country of Evil Barbarians who know nothing of probability theory or decision theory. There's a certain concept of "rationality" which says that the rationalists inevitably lose, because the Barbarians believe in a heavenly afterlife if they die in battle, while the rationalists would all individually prefer to stay out of harm's way. So the rationalist civilization is doomed; it is too elegant and civilized to fight the savage Barbarians... And then there's the idea that rationalists should be able to (a) solve group coordination problems, (b) care a lot about other people and (c) win...
Analysis of the gender imbalance that appears in "rationalist" communities, suggesting nine possible causes of the effect, and possible corresponding solutions.
I sometimes think of myself as being like the protagonist in a classic SF labyrinth story, wandering further and further into some alien artifact, trying to radio back a description of what I'm seeing, so that I can be followed. But what I'm finding is not just the Way, the thing that lies at the center of the labyrinth; it is also my Way, the path that I would take to come closer to the center, from whatever place I started out. And yet there is still a common thing we are all trying to find. We should be aware that others' shortest paths may not be the same as our own, but this is not the same as giving up the ability to judge or to share.
When subjects know about a bias or are warned about a bias, overcorrection is not unheard of as an experimental result. That's what makes a lot of cognitive subtasks so troublesome - you know you're biased but you're not sure how much, and if you keep tweaking you may overcorrect. The danger of underconfidence (overcorrecting for overconfidence) is that you pass up opportunities on which you could have been successful; not challenging difficult enough problems; losing forward momentum and adopting defensive postures; refusing to put the hypothesis of your inability to the test; losing enough hope of triumph to try hard enough to win. You should ask yourself "Does this way of thinking make me stronger, or weaker?"
Good online communities die primarily by refusing to defend themselves, and so it has been since the days of Eternal September. Anyone acculturated by academia knows that censorship is a very grave sin... in their walled gardens where it costs thousands and thousands of dollars to enter. A community with internal politics will treat any attempt to impose moderation as a coup attempt (since internal politics seem of far greater import than invading barbarians). In rationalist communities this is probably an instance of underconfidence - mildly competent moderators are probably quite trustworthy to wield the banhammer. On Less Wrong, the community is the moderator (via karma) and you will need to trust yourselves enough to wield the power and keep the garden clear.
I've developed primarily the art of epistemic rationality, in particular, the arts required for advanced cognitive reductionism... arts like distinguishing fake explanations from real ones and avoiding affective death spirals. There is much else that needs developing to create a craft of rationality - fighting akrasia; coordinating groups; teaching, training, verification, and becoming a proper experimental science; developing better introductory literature... And yet it seems to me that there is a beginning barrier to surpass before you can start creating high-quality craft of rationality, having to do with virtually everyone who tries to think lofty thoughts going instantly astray, or indeed even realizing that a craft of rationality exists and that you ought to be studying cognitive science literature to create it. It's my hope that my writings, as partial as they are, will serve to surpass this initial barrier. The rest I leave to you.
Knowledge of this heuristic might be useful in fighting akrasia.
Practical advice is genuinely much, much more useful when it's backed up by concrete experimental results, causal models that are actually true, or valid math that is validly interpreted. (Listed in increasing order of difficulty.) Stripping out the theories and giving the mere advice alone wouldn't have nearly the same impact or even the same message; and oddly enough, translating experiments and math into practical advice seems to be a rare niche activity relative to academia. If there's a distinctive LW style, this is it.
The fact that this final series was on the craft and the community seems to have delivered a push in something of the wrong direction, (a) steering toward conversation about conversation and (b) making present accomplishment pale in the light of grander dreams. Time to go back to practical advice and deep theories, then.
Generalization From One Example is a tendency to pay too much attention to the few anecdotal pieces of evidence you experienced, and model some general phenomenon based on them. This is a special case of availability bias, and the way in which the mistake unfolds is closely related to the correspondence bias and the hindsight bias.
Applied scenario about forming priors.
People don't actually remember much of what they know, they only remember how to find it, and the fact that there is something to find. Thus, it's important to know about what's known in various domains, even without knowing the content.
People are offended by grabs for status.