An Algorithmic Lucidity

"Empathy hurts.

"I'm grateful for being fantastically, unimaginably rich by world-historical standards—and I'm terrified of it being taken away. I feel bad for all the creatures in the past—and future?—who are stuck in a miserable Malthusian equilibrium.

"I simultaneously want to extend my circle of concern out to all sentient life, while personally feeling fear and revulsion towards anything slightly different from what I'm used to.

"Anna keeps telling me I have a skewed perspective on what constitutes a life worth living. I'm inclined to think that animals and poor people have a wretched not-worth-living existence, but perhaps they don't feel so sorry for themselves?—for the same reason that hypothetical transhumans might think my life has been wretched and not worth living, even while I think it's been pretty good on balance.

"But I'm haunted. After my recent ordeal in the psych ward, the part of me that talks described it as 'hellish.' But I was physically safe the entire time. If something so gentle as losing one night of sleep and being taken away from my usual environment was enough to get me to use the h-word, then what about all the actual suffering in the world? What hope is there for transhumanism, if the slightest perturbation sends us spiraling off into madness?

"The other week I was reading Julian Simon's book on overcoming depression; he wrote that depression arises from negative self-comparisons: comparing your current state to some hypothetical more positive state. But personal identity can't actually exist; time can't actually exist the way we think it does. If pain and suffering are bad when they're implemented in my skull, then they have to be bad when implemented elsewhere.

"Anna said that evolutionarily speaking, bad experiences are more intense than good ones because you can lose all your fitness in a short time period. But if 'the brain can't multiply' is a bias—if two bad things are twice as bad as one, no matter where they are in space and time, even if no one is capable of thinking that way—then so is 'the brain can't integrate': long periods of feeling pretty okay count for something, too.

"I'm not a negative utilitarian; I'm a preference utilitarian. I'm not a preference utilitarian; I'm a talking monkey with delusions of grandeur."

"I think Popper said something about the power of scientific reasoning: that our hypotheses can die instead of us."

"Neat trick, isn't it?"

"Except according to the third horn of the simulation argument, we are someone's hypothesis!"

Our subjective experience would have it that time "moves forward": the past is no longer, and the future is indeterminate and "hasn't happened yet." But it can't actually work that way: special relativity tells us that there's no absolute space of simultaneity; given two spacelike separated events, whether one happened "before" or "after" the other depends on where you are and how fast you're going. This leads us to a "block universe" view: our 3+1 dimensional universe, past, present, and future, simply exists, and the subjective arrow of time somehow arises from our perspective embedded within it.

Without knowing much in the way of physics or cognitive science myself, I can only wonder if there aren't still more confusions to dissolved, intuitions to be unlearned in the service of a more accurate understanding. We know things about the past from our memories and by observing documents; we might then say that memories and documents are forms of probabilistic evidence about another point in spacetime. But predictions about the future are also a form of probabilistic evidence about another point in spacetime. There's a sort of symmetry there, isn't there? Could we perhaps imagine that minds constructed differently from our own wouldn't perceive the same kind of arrow of time that we do?

There's this deeply uncomfortable tension between being an animal physiologically incapable of caring about anything other than what happens to me in the near future, and the knowledge of the terrifying symmetry that cannot be unseen: that my own suffering can't literally be more important, just because it's mine. You do some philosophy and decide that your sphere of moral concern should properly extend to all sentient life—whatever sentient turns out to mean—but life is built to survive at the expense of other life.

I want to say, "Why can't everyone just get along and be nice?"—but those are just English words that only make sense to other humans from my native culture, who share the cognitive machinery that generated them. The real world is made out of physics and game theory; my entire concept of "getting along and being nice" is the extremely specific, contingent result of the pattern of cooperation and conflict in my causal past: the billions of corpses on the way to Homo sapiens, the thousands of years of culture on the way to the early twenty-first century United States, the nonshared environmental noise on the way to me. Even if another animal would agree that pleasure is better than pain and peace is better than war, the real world has implementation details that we won't agree on, and the implementation details have to be settled somehow.

I console myself with the concept of decision-theoretic irrelevance: insofar as we construe the function of thought as to select actions, being upset about things that you can't affect is a waste of cognition. It doesn't help anyone for me to be upset about all the suffering in the world when I don't know how to alleviate it. Even in the face of moral and ontological uncertainty, there are still plenty of things-worth-doing. I will play positive-sum games, acquire skills, acquire resources, and use the resources to protect some of the things I care about, making the world slightly less terrible with me than without me. And if I'm left with the lingering intuition that there was supposed to be something else, some grand ideal more important than friendship and Pareto improvements ... I don't remember it anymore.

You hear people talk about positive (maximize pleasure) versus negative (minimize pain) utilitarianism, or average versus total utilitarianism, none of which seem very satisfactory. For example, average utilitarianism taken literally would suggest killing everyone but the happiest person, and total utilitarianism implies what Derek Parfit called the repugnant conclusion: that for any possible world with lots of happy people, the total utilitarian must prefer another possible world with many more people whose lives are just barely worth living.

But really, it shouldn't be that surprising that there's no simple, intuitively satisfying population ethics, because any actual preference ordering over possible worlds is going to have to make tradeoffs: how much pleasure and how much pain distributed across how many people's lives in what manner, what counts as a "person," &c.

"Don't worry, I'm sure it'll all go fine. Just be yourself."

"Of course. But ... how am I supposed to compute that?"

"You know secrets."

"Do I?"

"For example, you possess the True Secret of What It Feels Like to Be You, except you'll never tell, due to the best secret-keeping method of all, far stronger than any vow or oath ... namely, the lack of a language in which this secret can be expressed."

"Yeah."

"Actually, no, come to think of it, we might already have cryptography better than that. Unraveling the secrets of how the brain generates subjective experience is a monstrously difficult cognitive science problem, but it doesn't provably require unphysically large amounts of computation."

"I like adherents of Ideology X; I really do. Wonderful people with noble goals. It's just genuinely hard to communicate with most of them, because I assign a fairly high probability to hypotheses that they consider unthinkable—not even that; the problem with unthinkable hypotheses is that you can't consider them."

"You are mysterious."

"What, really?"

"Well, I can talk with you and guess what you're really like, but that's just me making inferences from your behavior; I could easily be horribly, horribly confused and wrong."

"So I'm not, like, unusually mysterious?"

"No, just ordinarily mysterious. That's still a huge amount of mystery. This is easier to notice when you have first-hand experience of being horribly, horribly confused and wrong about yourself."

"I somehow feel less bitter today."

"Yay!"

"Naturalistic explanations feel better than hatred; I don't feel outraged and betrayed that Safeway is not a Sacred Shining Beacon of Goodness and Food Provision unto the world or that Goldman Sachs is not a Sacred Shining Beacon of Goodness and Efficient Capital Allocation, so I shouldn't feel outraged and betrayed that college isn't a Sacred Shining Beacon of Goodness and True Education."

"Uh-huh."

"Even when Goldman Sachs causes a global financial system meltdown, I don't say, 'And that's why capitalism is Evil forever!'"

"Yes."

"Alternatively and isomorphically, maybe I should start saying—calmly, without malice or outrage, but just as a fact—that capitalism is Evil forever, with the caveat that I don't actually know how to fix anything myself."

"Why would you bother?"

"Because the insight that our current forms of social organization can't be the best possible, that you're actually allowed to think about them, even if innovations that actually work are hard, is really important and worth emphasizing and communicating. Because if you don't know and you don't think, you just get sucked in by whatever memetic attractor happens to be nearby, which is always suboptimal and often quite bad. I mean, you still get sucked in by whatever attractor happens to be nearby, by definition really, but it's nice to have some inkling of what's happening to you."

Dear reader, you occasionally hear people with conservative tendencies complain that the problem with Society today is that people lack personal responsibility: that the young and the poor need to take charge of themselves and stop mooching off their parents or the government: to shut up, do their homework, and get a job. I lack any sort of conservative tendency and would never say that sort of thing, but I would endorse a related-but-quite-distinct concept that I want to refer to using the same phrase personal responsibility, as long as it's clear from context that I don't mean it in the traditional, conservative way.

The problem with the traditional sense of personal responsibility is that it's not personal; it's an attempt to shame people into doing what the extant social order expects of them. I'm aware that that kind of social pressure often does serve useful purposes—but I think it's possible to do better. The local authorities really don't know everything; the moral rules and social norms you were raised with can actually be mistaken in all sorts of disastrous ways that no one warned you about. So I think people should strive to take personal responsibility for their own affairs not as a burdensome duty to Society, but because it will actually result in better outcomes, both for the individual in question, and for Society.

To the extent that good things actually get done in the world, it's almost always because somewhere along the line, a human mind designed that outcome. Valuable things typically don't grow on trees; fruit does, but in our modern civilization, it's because someone decided to plant that tree there, because a mind (not necessarily consciously, and not necessarily wisely, but all the same) generated a number possible plans, guessed at the probable consequences, and selected one to make real.

If you've been raised to believe that morality consists of obeying your parents or employers or schoolteachers, then it might seem like you don't have the right to make decisions, that decisionmaking (if it happens at all) is something done by committees of distant authority figures, not ordinary people like you, or that your right to decide things comes exclusively in the form of your right to vote, or buy consumer goods, or enroll in school courses. But it's not true; it can't be true; life itself does not emanate outwards from Washington or Wall Street or the Dean's office; with every thought and every breath, you are (implicitly, whether you realize it or not) controlling the tiny fragment of the world which is you, exerting an influence (however faint, however unpredictable) over the history of the universe outward and futureward from that tiny fragment.

Dear reader, I would not be so foolish as to advise you what you should do with this power, because I don't know anything about you or what situation you're in. I can't pretend to know what you should say or do, nor whom you should help, hurt, ignore, obey, or command. I just want to mention—because I wish someone had told me in so many words, this truth which still takes me an intense effort to keep salient—that it is a choice, that it is a responsibility. Not a responsibility to Society, or the moral law, or even to yourself. A responsibility over the outcome, whatever you make of that, or it.

Loopy is slang for "crazy", but I think it should be repurposed to refer to the quality of thinking the same sorts of thoughts over and over again, never breaking patterns, being stuck indefinitely at the same stage of intellectual development. You could argue that this is a form of craziness compared how an ideal agent would allocate cognitive resources, but I think it's pretty normal and common in our world, not the kind of craziness generally recognized as crazy.

We need a word that means almost the same thing as sellout (in the sense of "a person who compromises their principles for financial gain"), but conveys the idea that the problem is not selling out, but selling out for too low of a price. We all have to make trade-offs; there isn't any one principle that takes lexical priority over every other valuable thing in life: sometimes it makes sense to compromise your ideals in exchange for money or power or fame or fitting in.

But you could at least haggle!

"But it's kind of funny how my current idea of morality is so different and so much improved from what I picked up in childhood."

"Is it?"

"Well, funny is the wrong word; maybe it's better to say notable when what I really want is just to note it, just to make it salient, maybe eventually salient enough such that I can actually start to be moral for once, instead of continuing to sit and cry about how I was betrayed."

Hard Truths from Soft Cats opines that

Your flaws don't make you beautiful or unique. They make you flawed.

While I agree re beauty, technically, your flaws actually do make you unique: the number of ways in which one can be flawed is vastly larger than the number of ways in which one can be perfect; the probability that someone else would turn out to be damaged in exactly the same way you are, is negligible.

It's just that uniqueness is overrated.

What makes a true story inspirational? I think people usually use that word to describe happy stories, stories that make us think that the world is a better place than we previously thought. But sometimes I want to use it to describe sad stories that remind us that the world is far worse than just the parts of it we're used to seeing firsthand, stories about innocent people being hurt by arbitrary causes. It's inspirational in the sense of a call to action, a reminder that there's still important work to be done in the world: I can't solve this particular problem, but there's a reference class of people containing me (reasonably intelligent, reasonably ambitious people, striving to become more effective) who can help fix a reference class of problems including this one—and that is a sacred responsibility that must not be betrayed. Or however you translate folderol like "sacred responsibility" and "must not be betrayed" into something more basic (Bayes-ic?).

There are a lot of really important concepts that aren't easy to talk about, because we don't have standard words for them.

Like, there needs to be a word designating the skill or quality of possessing independent judgement—the ability to make decisions without getting distracted worrying about how to explain yourself to people who won't understand. Part of me wants to just call it sociopathy, but that's clearly not the right word.

The problem is endemic. Friend of the blog Mike Blume once lamented that we don't have a gender-neutral equivalent of gentlemanly. And we don't have an atheist equivalent of doing God's work, either.

Shut up! Shut up and leave me alone with my books!

I agree that it would be wildly out of character to learn JavaScript before Haskell, but life is not a story.

Oftentimes I initially want to write "It's not that X, but rather Y" to mean "You might think I'm implying that X is true, so I want to emphasize that X is false; Y is true, and that's what explains my position," but I worry that this idiom is ambiguous; someone is likely to interpret it as meaning, "X might be true, but it's not the relevant consideration; my position is actually explained by Y (which does not necessarily contradict X)", which doesn't mean the same thing.