Protect Your Slack, Delay Moloch: Why You SHOULD Defend Yourself With Artificial Rents

Inspired by Behold the Pale Child at Secretorum: "the arc of the universe is long, but it bends towards Bakkalon." (or Moloch. Moloch, at the bottom of the Darwinian/economic/political gravity well!)

The point of life is to be happy. How to go about this is mostly biologically determined. Yes, it's good to make others happy if you can, and to have making others happy make you happy as an incentive; for most of us, as social animals, this is also biologically determined. This position is that of a modern-day Epicurus, enhanced with and connected to facts about the natural world and our place in it. Not very controversial, you might think.

But I suspect that many people in the rationalist blogosphere will find it incredibly selfish to think first and foremost of oneself and ones own happiness, instead of the utilatarian (more specifically, Parfitian) long and wide view of everyone's happiness. (This more "selfish" position is not necessarily just individual hedonism, but rather would include having birthday parties for your kids instead of donating that money to dig a well in the developing world.[1]) In a curiously Calvinist-adjacent take, the implied position of the EA world (and tech capitalists telling young people their identify comes from working unhealthy hours and sacrificing the rest of their lives) is that you should de-emphasize your own happiness here and now since it's such a small drop in the ocean of possible conscious experience throughout time.

And yet if people are biologically limited by the link between their happiness and the amount of work they can do - and the kind they can do - and for whom they do it - and they are - what you're asking is many people to sacrifice their own happiness for an uncertain outcome, for an uncertain philosophical position.

The position of working 110% all-out all the time is not just something from the world of Effective Altruism (EA), etc. In a recent post on the Slatestarcodex (SSC) subreddit, in a discussion about the Musk-like approach to constantly fastforwarding everything and having work be eternal crunchtime - a commenter stated that once in a great while, such a push was okay, but it's not sustainable. I would go one step further: I want to enjoy my life, working hard diminishes that, focusing on any one thing to the exclusion of most others diminishes that, and you should avoid crunchtime and working hard wherever possible. (That is, I value slack - see Church of the Subgenius - and I will defend that slack if necessary, even if I have to do it surreptitiously.) Wanting to enjoy your life, and do more things you directly enjoy more often, and fewer instrumental things, is not something to be ashamed of. That's why I'm posting it online and telling you it's good to feel the same way.[2]

It's true that if everyone thought this way, then life-saving and -improving technologies would progress much more slowly. But herein I'm taking the (apparently very hard to grasp position) that I neither want to work that hard, nor do I want to get in the way of people who do want to work hard. I say in all seriousness: good for them, I'm glad we have people built this way![3] But don't feel bad if you're not one of them, and you're almost certainly not. I'm definitely not, and I feel great about it! I've even turned down promotions for this reason. Again, not controversial, I wouldn't think. But it feels very much like an emperor's new clothes position to take.

The opposite of slack is hypercompetition, which I don't have to further describe to anyone living in the developed world in the 21st century, and I would argue a big part of Moloch is hypercompetitiveness (Moloch in the sense of Scott Alexander's synechdoche for a self-perpetuating system with serious and unintentional consequences that benefit nobody.) There is only so much work you can do; you need some slack, and though our modern Molochian culture has trained us to hide our slack-seeking from ourselves, we do it, or we burn out. And part of the outrage at people finding ways to guard slack is a result of hiding our desire for slack from ourselves (read: reaction formation and the predictable reaction to seeing others fulfill their fantasy) when a source of slack protection is noticed. (See: "tears of rent-seekers" regarding taxis, academia, government, or any other area where people have goodness forbid given themselves some extra slack to help them enjoy their lives.)

Other strategies: shrouding - which normally means companies trying to avoid competing on price by making their pricing opaque, but works in the labor market too when workers cooperate to obscure measurement of output - outlawing payment for piece work was a major victory. Another: avoidance of direct market exposure, or any situation where your livelihood rests on your having to react in real time to developments - usually the more layers within an organization away from a customer interaction surface or competition with other organizations, the quieter your life is. (Must be balanced against the risk of paroxysmal collapses; the cycle-time of this class of org is relevant to your choices here (nations = centuries, companies = years or decades if already long-established.)

Some concrete examples are in order, of how you can and should protect slack and benefit your life by erecting artificial rent structures.


SITUATION 1 You're the leader of Organization A. You believe in what the organization is doing, genuinely care about the people there and want them to have good lives, and as a result you "leave some money on the table" by not expecting them to work that hard or otherwise sacrifice their well-being to the organization, as long as they keep the wheels turning.

Then Organization X comes along (for the Parfit-style calculators out there, let's say it has the same number of people), which does NOT care about its people this way, and they are constantly sacrificing themselves, or at least on a sort of psychological Malthusian frontier (of stress rather than starvation.) This might well be an Elon Musk company. Organization X eats Organization A's lunch, and Organization A is destroyed or absorbed, along with the lifestyle of the people in it.


SITUATION 2 Same as above, but you're the leader of Organization B. You know it is likely that if you do NOT drive your people to self-sacrifice, then a Muskite will drive theirs in such a way, and then they'll come for you. So for your organization to continue existing, you have to work them to the point of self-sacrifice. You do this, and keep existing, but the people who work at your organization are now miserable.


SITUATION 3 You're the leader of Organization A, same as Situation 1. Except you have a plan. You want your employees to have a good life but you know that the Muskite misery engines out there like Organization X will come get you. So you make a couple calls to a governor or legislator, take them golfing and make some arrangements, etc. Organization X now finds you have an administrative or legal moat - an artificial rent protector - for example, to do what your org does they have to be in a certain consotrium and no one will let the Muskite org join, or Organization X can't operate in a certain business in a certain territory, unless the workers within Org X get lots of protections. You know this can't work forever, but it will work for a while, and benefit the people you care most about. Organization X loses its advantage in being willing to essentially trade personal slack for victory. People on SSC read about this, and cry their eyes out talking about Rents, and how you're immoral for depriving the rest of the world of the fruits of your labors (invisible tragedies, etc.)


I used to join in with the "ha ha, rentiers dying, suck it taxi drivers" until I realized that within a few years, AI will be able to do all of our jobs, and the value of labor will race to zero. Of the strategies I've mentioned, only legal artificial rent structures have any chance of lasting for any length of time. So I'm unashamed to admit I would rather work for Organization A in Situation 3, and unless you're the 1% of the 1% in productivity, you would too. (I hate to be the one to tell you, but if you think you are a 10x 1% of 1% superstar, you are much more likely to be delusional than an actual superstar, and the angrier that statement made you, the more likely you're delusional.) Of course, sometimes the rents come "honestly" from an innovation - but then again, even patent protection is an artifical rent, since it's not just the innovation itself. Mostly rents come "artificially" from barriers like the ones I've described. Taxi medallions, medical licenses, etc. although in most cases there's usually at least some non-bullshit reason for the certification, or guild membership, whatever it is (e.g. it's a quality signal.)

Note that I've written these thought experiments with you in the position of the leader. But you're almost certainly not. If, in a true Rawlsian approach - if you fell out of the sky at random into these thought experiments - you'd probably be a rank-and-file employee. In that spirit:


SITUATION 4 You're an employee (not the leader) of Organization A. You believe in what you do and what the organization stands for. Your leader seems to genuinely want everyone to have good lives and doesn't work anyone too hard. As you smirk and murmur to your colleagues at pool parties, this is because the leader is friends with the governor, and got a law passed artificially protecting you from competition, which is why you have a good income without working too hard.

Then the leader dies or steps down, and a new CEO takes over - one who reads SSC and Marginal Revolution. "Enough with this laziness! Company X has their own lobbyists, and we can't wait for them to get the law repealed and be caught off-guard. 80 hour weeks! No vacation or weekends if you want to be considered serious around here! Constant aggressive deadlines! Do it 10x faster! We're depriving the rest of the world and future generations of the fruits of our labor, how selfish that is, think of all the hidden tradegies! Don't like it? Emigrate/quit and go to our competitor, who will probably have to do the same thing to keep up anyway." Would you say "Yes! Finally, our new leader is high-agency, and this is the moral thing to do instead of collecting rents"? Yeah, sure you would.[4] If you do, you burn out, ruining your health and family life, plus you have no more time to read SSC.


Certainly it's a difficult balance to find, and often you're just surfing a temporary inefficiency wave until that wave breaks and you're back in the same Molochian world as everyone else - but you should try to find it and ride it as long as you can. In the long run, we're all dead anyway. If you can have 5 or 10 more years of slack instead of zero more years, you are not being immoral to take it, and (for the Parfitians in the back) you can't be sure that the only thing you'd do by missing out on the slack is making yourself miserable with no other impact, thus doing the immoral thing of increasing the suffering of the universe on net.


[1] I've noticed that the tech world in general and EA especially is a haven for those who in the abstract, are horrified at the existence of slack (or at least that's the non-revealed preference.) In general consequentialists tend to neglect deontology - the role of duties in what decisions are moral. Consequentialists tend to look for abstract principles for actions to adhere to, but actions are not disembodied principles, they occur in time, and space, and social space - that is, in the context of whatever history and relationships, if any, you have with the people affected. Deontology clears up a lot of the confusion about what to actually do and when to do it, and who to do it with/for. I've also noticed conscientious younger people tend to be consequentialists, and older people season their outlook with more deontology as they age.

[2]Maybe this whole essay is just my own psychotherapy, justifying the following to myself: as a physician, every time I go home at the end of the day or take a day off, I am depriving people of potentially life-saving treatment. Some physicians, more in previous decades than today, kept this in mind and worked ridiculous hours; many modern healthcare organizations are more than happy to take advantage of this mentality of self-sacrifice to make another cent, and then when you start making mistakes because you burn out, they kick you to the curb. Not unique to medicine of course, but I'm very comfortable protecting my time so I can have slack and enjoy my life, and what's more, I limit my responsibilities to my established patients, and not some abstraction of "possible humans in the universe". If you're a naive consequentialist (who doesn't understand deontology or respect the limits set by biology) you've probably dismissed me as Jeffrey Dahmer by this point.

[3]To beat a dead horse: this is not an anti-hard-work screed. If you like to work hard, focus on one thing and one thing only, you find it rewarding, great! Part of civilization's success is that we've set up a system that rewards you, and where the rest of us also by diffusion get the benefit of the wealth and technologies you create. But if your choices start taking away my slack - I'll ask my guild to take our Congressman golfing, after which an artificial moat may mysteriously appear. For a relevant culture-wide take on the same: I once read an account of an American traveler in Japan who said it's great to be a foreigner in Japan - because it's a safe, clean, beautiful, quiet place, due to the crushing social obligations of Japanese culture that keep it this way, and as a foreigner you can free ride on this. But you obviously shouldn't do anything to make it harder to keep the country that way!

[4] SSC surveys have consistently shown that oldest siblings are more likely to be readers. Though it's a stretch, it does make me wonder if an oldest-sibling-rich group concerned about these topics might tend to lack a healthy level of resource anxiety (no older siblings to finish all the dessert before you, hog the TV or soccer ball, etc.) This would lead them to always assume that protecting slack can only be about stupidity or laziness - "Aw, we ALWAYS have to stay on the little kid playground because of them!"

Lists of "Bizarre Beliefs" Reveal Difference Between True Belief, and Tribal Team Cheers

tl;dr Many truth claims - beliefs - are actually just tribal team cheers, or emotional signals, with the propositional verbal content merely superficial. We can get confused and react to these as if they are truth claims, especially because they people saying these things insist that they are. We need a name to distinguish them from real propositions - let's call them dolphin beliefs, because of their superficial similarity to true, propositional "fish" beliefs.


You should read Aaron Bergman's review of Fantasyland, a book about American's relationship with conspiracy and magical thinking, today and over the decades. He cites surveys which show, for example, that one in nine Americans believe they have seen the devil driven out of someone. Others he cites are about Obama being born in Kenya, vaccines causing autism, and ghosts. Recognizing that no one is immune to irrational beliefs, Bergman identifies what he thinks are his most "fringe" beliefs. And here I also engage in this exercise, not because I think you're particularly concerned with my fringe beliefs, but because it's interesting to see the differences in his and my list, vs the kinds of things discussed in a book about American conspiracy thinking.*

A few of my own bizarre beliefs:

• Panpsychism - consciousness is a fundamental property of the universe itself.
• There are living cells on Venus which explain the unknown absorbers and presence of phosphine. (See below for an update to this as of 2025.)
• As we explore the solar system, we will find evidence of von Neumann probes on asteroids.
• There will be nuclear weapons used in war in the lifetimes of many readers, and we don't talk about this nearly enough.

There are two characeristics to note about these fringe beliefs - one of which Bergman and I share with the conspiracy-believers cited in the book, and one which I think we do not.

1. We are not good at knowing what will seem strange to others.
   
   
2. These beliefs are not central to identity.

I think if you asked the devil-drivers to name their fringe beliefs, they would (in keeping with #1 above) not necessarily realize that devil-driving is seen by many others as a strange, fringe belief. Similarly, when voluntarily producing a list like this, I probably haven't been able to identify the beliefs I hold that would most shock most readers. This occurs because we're all embedded in communities. Devil-drivers know a lot of other devil-drivers; the belief doesn't seem strange in that context.

As for #2 - I can't speak for Bergman but I know that, if I encounter a strong argument against panpsychism, or data from a probe in the Venus clouds showing a completely mundane abiotic process that produces phosphine, not only would I probably change my mind - I would not become hostile and defensive, as if I were being personally attacked. Resistant, disappointed, a bit embarrassed to have been wrong in public - sure. But not angry. Whereas I think if you were to engage a devil driver and explain why their belief may be wrong, I predict they would become hostile and defensive, as if they were being personally attacked. Same for antivaxxers and birthers.

This underlines the core difference in two types of beliefs. There are actual hypotheses - what a belief ideally always is, able to be updated by new information – and then there are the tribal team cheers of religion, politics, or conspiracy communities.

If we think of beliefs as a good materialist should, we think about what is actually going on in the nervous systems, and how the behavior of the organism differs systematically in a way that can be categorized or at least placed on a spectrum. Notice that it's not merely isolated "trapped priors" we're dealing with here - antivaxxers and devil drivers don't just calmly reject arguments and information and continue to believe what they already believed. There is community, identity, and emotion involved.

I therefore think we should consider whether the "beliefs" of devil-drivers and antivaxxers are truth claims at all, or something else.** At the very least we should consider whether their utility is more as tribal team cheers than as truth claims.

The implication here is that the superficial content of the belief is not the only determinant of whether it is a functional, updateable belief (a hypothesis) or a tribal team cheer. For example: say I learn that there is going to be a meeting of a local club to discuss the phosphine and unknown absorbers in the Venusian atmosphere. Excited to talk about it with like-minded people, I attend. At the meeting I find people talking about how they just know in their hearts there is life on Venus, that NASA is trying to hide the evidence, and that they don't care what additional evidence the probes might find. In fact when I suggest we send more probes they are actively hostile!*** Whereas the club members and I would both say "There is life in the Venusian atmosphere", I have a hypothesis, they have a tribal team cheer, though the superficially the content of the claim is the same. (The hypothesis IS just the content of the claim; the tribal team cheer is a cake of social behaviors with the words of the truth claim as icing.)

(Added later: as of 2025 I have updated my belief and accepted that the unknown absorbers are not life. Not only was that phosphine paper a product of bad spectrometry that failed multiple attempts at replication, there are at least two papers - Jiang et al 2024, and Egan et al 2025 - that have advanced good candidate abiotic explanations for what the absorbers could be. Needless to say I am now a pariah in the Venusian life community, who say I was never actually a true believer, etc.)

In fact, focusing on the process of belief, rather than the content of the belief itself, is what we do in psychiatry. If someone is convinced his wife is cheating on him with absolutely zero evidence, even if she confides "actually I did have a drunken one-night stand ten years ago but he doesn't know about it" - that's still a jealous delusion. He doesn't have a good reason to believe it. The Venus club's stated belief is a community and identity device, not a cognitive tool for explaining the world. Hence bizarre statements, in the rare occasion when they are discussing it with people from outside their community, like "I just feel that it's true", "this is offensive", and "this is a personal attack."

Because it's easy to be confused by tribal team cheers which do indeed look like truth claims, especially when the tribal team cheer-ers are loathe to admit that it's not really a truth claim, it's worth identifying the tribal cheers as something different from hypotheses.

You're probably familiar with the idea of a shibboleth. For me, the belief in Venusian life is a hypothesis; for the club, it's a shibboleth - or at least, much more of a shibboleth than a hypothesis. The more of these characteristics it has, the more likely a belief is a shibboleth than a hypothesis:

• Avoidance of any testing
• Anger at questions, as if somehow being personally attacked
• Formation of identity around the belief
• Reason for belief is emotional
• Association with community around the belief

Devil-driving, birtherism, and antivaxxer-ism are shibboleths. Panpsychism is a hypothesis. In the future of epistemology, people may be amused but charitable that we did not make this distinction, just as we think of people five centuries ago who didn’t understand that dolphins are not fish. For that reason instead of calling these types of beliefs shibboleths and hypotheses, let's call them dolphins and fish respecitvely, to emphasize their superficial similarity, and because many dolphin beliefs are actually not in-group team cheers, they're just used by individuals to send emotional signals.


*It's worth pointing out that the types of beliefs we articulate, when asked what our most surprising beliefs are, are generally about the external world, not internal beliefs like "I'm unlovable" or "I can't accomplish important things" - even if we're frequently aware of such beliefs, we guard them closely. I think this is more likely out of fear of the impact on others' opinions of us, rather than a shrewd calculation about what people want to hear about.

**At one point there was a debate in psychiatry as to whether delusions are really beliefs. My argument is that they are indeed something neurologically and behaviorally different, though this is an academic or semantic distinction at this point.

***Compare to eg creationists, who often spend much more time talking about how their enemies are suppressing them than providing actual arguments and data, making predictions or trying to do something pragmatic and useful with their "theory". Where are the creationist biomedical companies?

Addiction Has Three Types: To Pleasure, Flow, or Meaning

Positive psychology research models happiness as reducing to three components: pleasure (chocolate, sunsets and orgasm), flow (losing yourself in an activity; "action meditation") and meaning - feelings of value and connection and identity within a community. My fellow Americans and I have a tendency to reduce the pursuit of happiness to the pursuit of pleasure. Pleasure is important, but it's not the whole game, and modern consumer society may have specialized in that component and lost the other two.

We commonly think of addiction as a problem of pleasure and pain: pleasure when you consume the addictive substance (or perform the addictive activity), and suffering when you do not, after you're hooked. It's not hard to see how it's not just meth, but sex or food, that could be the subject of addictive behavior. Humans have continued to get better at creating goods and services that cause repetitive behavior in their targets - think brand loyalty, processed drugs (think coca leaves to cocaine; increasing the sugar in everything.) We have probably become better at identifying avoiding these addictive goods and services. Still, I'd bet good money that over time, they're becoming more addictive faster than we're getting better at resisting them.

Through the lens of positive psychology, this starts to look too narrow. It's not just the pleasure component of happiness that has been exploited. Addiction to flow and meaning exist too.

Probably the best example of addiction to flow is video games. They are designed for this purpose, and there is evidence that they are damaging the productivity of young people, particularly young males.

Then there is community addiction. The most harmless form is what Facebook exploits to get you to keep checking whether your friends have liked your post. The more concerning form is that of religious cults, or small ingroup-vs-outgroup communities (often based on unearned, non-opt-outable qualities like race or religion.) This is actually the one that has the most potential for harm. Postwar Japan and late 60s America both featured a shock to the automatic meaning-generating aspect of national community. This is happening in a second wave in the West. It pains me as a staunch atheist to say this, but it's becoming clearer that the disappearance of religion as a community-builder has not been a boon for everyone. If you're also an atheist and that sentence made you squirm, then here's a thought: a substantial part, maybe even the majority, of young American males who support Donald Trump are not religious.

I have enough confidence that these three types are legitimate subtypes that I think it would be appropriate to have them DSM as specifiers. Indeed, the discussion over video game addiction has been ongoing for some time.

For the Rationalist Community: More on Why Choice of College Doesn't Matter That Much

After I wrote the last post (telling people college choice doesn't actually matter much in terms of happiness in life) it struck me how directly opposed it was to the message that's been growing over the last few years in the rationality community, much of which overlaps with people interested in existential risk (x-risk) and effective altruism. Since Scott Alexander is kind enough to put me on the blogroll at Slatestarcodex, I occasionally get readers from there. Some of them are in high school and are deciding what to do with their lives; some are parents, helping their kids decide.

(An interesting take on perils to be avoided in college culture from Dan Wang)

A good starting point is this SSC post, which notes: "In general, [Dale & Krueger 2011] find no advantage from attending more selective colleges. Although Harvard students earn much more than University of Podunk students, this is entirely explained by Harvard only accepting the highest-ability people. Conditional on a given level of ability, people do not earn more money by going to more selective colleges...Overall, unless people come from a disadvantaged background, there’s surprisingly little evidence that going to a good college as an undergraduate is helpful in the long term – except possibly for a few positions like President or Supreme Court justice." No argument that many colleges are getting harder to get into; but clearly, it doesn't matter as much as you might think.

And that's the point I'd like to expand on. There are two messages that we're getting these days from the ambient culture about college decisions:

• If you don't get into the college of your choice, your life is ruined, that's your one chance, you're now a loser and you won't be able to get a good job (society's current message)
• In order to work on anything meaningful, it has to be something involving AI, x-risk, or EA/rationality-adjacent topics. (this community's implied message)

It turns out that neither of these have much to do with meaning or personal happiness.[1]

I worry that earnest young rationalists, who are already anxious enough when they recognize the pervasive irrationality in society, will have their anxiety exacerbated by these discussions and the extra fuel they add to the (unnecessary) college admissions madness I'm arguing against here. There is also something to be said here against the tendency of people in the rationalist community to define themselves based on intelligence. Intelligence is only a proxy indicator for actual accomplishment. (No one was ever recruited onto a basketball team for being 6'10". They're recruited because they can play good basketball. Yes, height is correlated with that, but it's still only a proxy.) And you have to keep in mind, getting into college isn't the accomplishment.

There are certainly choices you can make that will damage your happiness in life. These do not have to do with which college you go to. They are things like abusing substances, driving unsafely, harming the people you care about or allowing them to harm you. I should clarify that this is not an argument that college is pointless, that you shouldn't try to get in, or that choice of program is 100% entirely irrelevant to your life. I think you should try to get into the best school you can, that's best-suited to what you want to do. What you should take away from this is that there's no point killing yourself to do it, and don't worry too much, because if you don't get in to your top choice(s), you'll be fine anyway.

The overall message is that wherever you go to college, it will not significantly harm or help your chance at a good life.

College admissions madness – noticeably worse in the 30ish years since I applied to undergrad – is mostly motivated by status considerations. Yes, status helps you do other things, much like developing a chess piece's position, and I've wondered myself about how much time to spend looking impressive vs actually "doing things." I don't have an answer for that, other than for me personally, the answer is clearly not "indefinitely increase status throughout life." I tried it. It makes me miserable and is unsustainable, and defeats the purpose of being able to use the status to "do things." Some problems are 1) that we can never know when your quest for status is actually just motivated by the immediately positive feedback you get right now from someone recognizing your status (rather than it being a vehicle to accomplish other things) and 2) that status is an ordinal good – it's about ranking, not absolute value. Moreover, it's necessarily zero sum. If you rise one spot in a status competition, someone else drops one spot. Smart people – emotionally and intellectually smart people – stay out of zero-sum games, including status competitions. One major drawback of the internet is that suddenly we're all aware of the same status hierarchies and aware of how the other half live, and we are drawn into the game, instead of appreciating that in absolute terms we mostly have good lives, and will have the same life regardless of where we rank in status competitions, including "where was your undergrad institution ranked?"

It's tragic that, so far as we know, this is the one place and time in the universe where such a thing as happiness can exist, and the most evolved beings on this planet who live in wealthy free nations with material lives our ancestors could only dream about, are squandering our happiness to worry about whether their neighbor might think they're better.

So here are some follow-up questions and objections that I imagine might be voiced by a young person connected to the rationality community, puzzled by what they're reading here, against the backdrop of other messages they've been getting.


"I want to work on the most important problems, especially x-risk problems. I can't do that unless I get into these institutions."

It is very unlikely that getting into a certain university is 100% gating to whether you can do anything that contributes in a certain area. (Worth asking here: are you working on nuclear x-risk? Why not? Even the risk vs probability graph here shows nuclear war having an expected value worse than rogue AI on the 200 year horizon; I don't think the probability of nuclear war drops as much with shorter time periods as rogue AI drops. Seriously. Work on nuclear weapons reduction, that has a much clearer payoff.) If you don't end up in a position to do anything related, you'll still have a good life. And to that someone might say, "So you're admitting you're just interested in having a good time, and you're not stressing too much about working on x-risk or any other cause? How incredibly selfish!" This is the paradox. If you want to work on x-risk, and you apply and can't get in to your preferred programs – guess what? It's probably because people better than you at those skills got the spots you were applying for. And they're going to do a better job working on those problems than you would have, so the work gets done anyway. So there's the paradox – unless you're saying what matters is you and your little ego getting into that spot, rather than progress being made, you shouldn't be upset about that.[2] It's fine, and you'll still have the same quality of life you would otherwise have.

If you're objecting to my statement that the admissions process is fair and accurate, and you're saying "the applicants who beat me out won't necessarily be better than me – they're just people who hire better essay writers, or whose richer family gave more money to the university, or who are better looking, or [endless list of unfair gamey things that have nothing to do with success in the chosen field of endeavor.]" And you are quite possibly correct! So this doesn't reflect on who you are, you have no control over it, and you'll still have a good life!


"But I can't do [exact career] unless I get into [chosen institution.]"

Rarely are career tracks so black and white. Yes, it might be harder to go from A to B (become a professor, get funded by a VC, etc) if you don't start out at a top institution, but rarely is it a go/no-go situation. But even that difference in access is not really about the education, and is more about networking and access. You can do that with internships, visiting your friend who did get in and making contacts, or doing projects with contacts you make online. If you're hoping you can just rely on the credential you get from institution X and are not doing anything related to that career before you get there, outside of the institution...then maybe it's just the status of the institution you want, and not what it can actually do for you in terms of making you better at that field of endeavor? (See more on this below in the "But I won't be able to learn..." section.)

In fact there are very few one-hour periods in your life that strongly determine the rest of your life – life is mostly the accumulation of small decisions made multiple times a day. The discrete future-determining events tend to be either not programmed (e.g. random introduction at a social event) or have nothing to do with your career (asking someone on impulse for their phone number, and later you start a family with them.) You might object that there was an admissions test you didn't do well on, but guess what - your ability at that test was determined well before you sat down, in fact mostly before you were born.[4] And even then, a bad score because you were sick doesn't destroy the rest of your life the way, say, mistreating a family member might.

An anecdote, with thrills and spills: a med student that I once supervised as a resident got back her Step 1 score while I was working with her. For non-medical people, this is the single most important test score that determines how competitive you are as a residency applicant – both in terms of the programs you might go to, or even the specialty you might get into (some specialties are harder to get into than others.) She remains the best med student I ever worked with – but when she got back her score it was quite disappointing to her. "I feel like my future just closed off! I can't get into [competitive specialty.]" And with her score, she was probably right (a spill.) So I asked her to go for a walk with me and I told her I thought she would have wasted her talent on [competitive specialty] and that she'd be a talented psychiatrist (which is not as hard to get into, though this is changing – fashions in medicine are funny things.) Fast forward, she is now in her psychiatry residency (thrill), and got into [extremely competitive program.] I told her to stay at our program, partly out of selfishness at wanting to retain her and work with her (which I freely admitted) but I also thought she would be happier. I haven't spoken to her recently but I've heard that she hasn't enjoyed her training experience so far, for deeply unpleasant reasons that have to do with the culture of the place but not the rigor or work hours (a spill.)


"If I want to be a professor at a major university, or a physician, or a lawyer, there are big differences in my chances based on what college I'm applying from. There are real hoops and gatekeepers."

First, see "But I can't do exact career" above. Also, the same thing applies as with X-risk above: if you're really in it for patient care, or justice, etc. and you didn't get in, then someone else better than you got your spot. Fine! Your patients or clients will still be taken care of, and you'll still be happy as not a doctor or lawyer. (In fact, likely happier. The burn-out statistics for these professions are abysmal. And sometime read up on the lives of people who had physics units, or other equally Important Things, named after them. Maybe you're more selfless than most people – actually probably not – and most people would rather have a good life than a unit.)[5] Regarding getting a PhD and getting a full-time tenured faculty position, in a place you might actually want to live for its own sake – you're entering a lottery similar to that jock you scoffed at in high school, who said he was going to play pro sports.

Also, if you can't get into a good institution, maybe you're not cut out to be a doctor or lawyer. That's fine, and in fact I'm happy for you that you didn't get stuck in those careers, where unless you're suited for them you would be abjectly miserable. (See below, "There's a clear hierarchy...") Even within medicine, I meet psychiatrists who hate being psychiatrists, surgeons who hate doing surgery, etc. – and they have bad lives. We live in a technologically advanced wealthy country. There are many, many, many things to do with your life that will be fulfilling, make money, and give you a chance to create value for others. I once overheard a middle-aged mountain guide on Mt. Shasta once talking with the people in his climbing party, and it turned out he had applied to med school and then withdrawn his application. He was having a blast on the mountain that day and showed no sign of regretting his life choices.


"I see a lot of information online about how VCs and top companies only recruit young people from elite universities. And various people who talk a lot online about start-ups and technology companies say that unless you're working in technology, you're not doing anything important."

If CNN ran stories all the time that said, really the only significant thing in your life is to become a TV journalist, and it's incredibly competitive and you should start thinking NOW about where to attend so you can get hired...I hope you would laugh at the obvious self-servingness of it. So when you read online about how tech is the only thing that matters – how is that different?

It's worth pointing out that VC partners like young founders, despite data showing that older founders are more likely to be successful. That is self-evidently because they can control the founders more easily, and make them work harder – no families, no expectations of a healthy life-balance. (That little twinge you just felt that you must be a loser if you want time with your family and a balanced life? That means that "the 1%" have already done their job on you.) In the same way, consulting firms are not interested in older applicants, because they can't mold them and get them to work stupidly long hours – a fully formed personality and set of values is not useful when you want to control someone. So don't be so eager to put yourself in this position.


"There's a clear hierarchy in life. Don't kid yourself. This is all just wishful thinking."

There are certainly status hierarchies. If you haven't bought into their lack of innate value by now, you should stop reading. But it's true, we're human, and they affect us – which is why the healthiest societies are ones with multiple overlapping status hierarchies, and the healthiest people are members of multiple status hierarchies. First, there's not just one dimension with "MD/PhD at UCSF" (in my world) at the top, and everyone else below that. Whatever is at the top of the hierarchies you spend your time worrying about, rest assured that the majority of human beings don't care about it. It can be easy to forget that other status hierarchies exist when you're inside one, but if you can, avoid taking them seriously, and stop worrying about who's above and below you.[3] (If there is a secret to a happy life, it's the ability to do this. I'm still working on it myself but when I can do it, it feels great.)

Second, to the extent that you're willing to subject yourself to a status hierarchy because there are good people participating in it who you respect and whose opinion you admire – make sure it's one that fits you. People differ tremendously in their innate talents and temperaments. You might not be cut out for a STEM field. You might not even be college material. There are many many many things to do in life that are fun, create value for others, and can make you a decent living, that don't require a bachelor's degree (ie trades.) Don't rob yourself of a career you'd be great at because classmates look down their noses at your choice. There are many many many people stuck in postdocs or at big law firms wondering why they chose this path. Ask yourself seriously how much you're collecting status coupons (degrees from prestigious institutions) to reassure yourself that random strangers who don't matter to your life or goals will think you're impressive, versus how much those status coupons are actually getting you toward your goals. If you don't know what those goals are, figure them out.


"But if I don't get into a good STEM program, I won't get a job."

Most jobs that require a bachelor's degree, still do not specifically require STEM degrees. With the exception of programmers, we produce way too many STEM graduates, particularly ones who want to go beyond undergrad. (See "If I want to be a professor" above.) You can learn programming from any institution and you'll be in demand. You can also learn programming on your own (see "But I won't learn" below.) If you say you want to be a programmer but your reason for not spending more time learning it is that you didn't get into XYZ University...chances are, your self-starting passion would not have been kindled there either.


"But unless I get into XYZ University, I won't be able to learn [field.]"

Don't MIT and Harvard literally have all their courses online? Why aren't you already taking them? If you need a professor to threaten you with bad grades to make you learn the material, you are not a talented and self-driven achiever in that field, and going to such a program will give you a piece of paper that impresses people and maybe gets you paid a little more, but will not turn you into a talented and self-driven achiever.


"Fine, but that 10% income premium does make a difference over time. If I don't get into an elite school, I can't join the 1%."

First: if you're thinking your salary is going to be what gets you into the 1%, then you're woefully middle class and don't understand the 1% at all. Being rich is about capital, not salary. I'm a physician, but I still rely on my salary. I'm not elite. I have to go to work on Monday or I can't pay my mortgage. So your university experience is very unlikely to get you into the 1%. Second, above about US$70,000 (maybe a little more adjusted for today's dollars and your local cost of living) the happiness curve plateaus. One caveat – if you're the first generation in your family to go to college, or you come from an underrepresented minority or immigrant population, college does give a bigger benefit to your future prospects than others, and the better the college, the more disproportionately bigger the increase.


"Sure, it's easy for you to say 'don't worry about getting in, status doesn't matter.' You got into med school and you're now established in your career."

Oddly enough, the thing that I think has most made me relax about my career is having a family. It really does put things in perspective. So it's not really "I did it, so it can't be hard", it's more "I realized my career is actually not as important as I used to think it was." And the reality is that a job is a job (see "Fine, but that 10% income premium" above.)


"Status does matter. I can't help noticing how people regard me. So I'll be scared and depressed if I follow your advice, and I'm still going to try to get into the best school possible."

First, you should try to get into the best school possible, but you shouldn't kill yourself trying, and you shouldn't feel bad if it doesn't work out. But if you feel controlled by your own need for status recognition – this actually IS something you have some control over (try CBT!) Second, there are two sides to the coin. What if you go to Harvard, and don't do much with your life? Many Harvard grads have regular jobs, and are constantly feeling pressured from other Harvard grads, their coworkers, families, etc. when someone says "So you went to Harvard, and now you're doing...this?" (i.e., the same job with same office that someone who went to State U is doing.) I have known Harvard grads who feel this, and it's unpleasant. They have the same life that a State U grad has - which is fine - but also the feeling that they've disappointed people - which is not fine.

An anecdote: an acquaintance from my undergrad (a decent but not awesome public university) was interviewing for a prestigious international scholarship. When they asked him what he would do if he didn't get it, he said, "Oh I'll be fine, I already got [other, less prestigious international scholarship.]" (Which he had.) He soon realized this was the wrong move, and needless to say did not get the scholarship he was interviewing for. Guess what? He has had and continues to have an amazing career, so much so that I'm changing details and not naming him because you might be able to figure out who I'm talking about based on his presence online. And, it's very difficult to see how he would be happier OR more successful if he had gotten the "better" scholarship. It is also difficult to believe that he would not have had a fulfilling career even if he did not get any prestigious scholarships (and again, note that successful is not the same as happy!) And looking at top scholarships - Rhodes Scholars are smart people, but they don't change the world. The people who get into top scholarships and programs are the people who are best suited for getting into top scholarships and programs - not necessarily doing the things that those programs are ostensibly training people for.


"I was reading your other post, and I read where you say you looked for data on happiness outcomes from attending universities of varying quality – and there's little to no data on this? This is even worse than I thought! How am I supposed to make a decision! I'm flying blind in my quest to have a good life!"

No you're not. What data we do have suggests your quest for happiness is minimally impacted by where you go to university.


"This is so selfish. If everybody thought like this (maximizing personal happiness) then progress would grind to a halt, the AIs will get us, etc."

But guess what! Not everyone thinks like this. But, if you're willing to sacrifice your personal happiness, I truly do appreciate your decision. I just want you to realize that's what you're getting into. Working 20 hour days will not make you happy, even if you get into the school/company/etc. of your dreams.


"Look, I know I will be happy if I get into [college program of my choice.] You can't explain that away by saying my life will be fine either way."

Yes, you will have a bump up in your mood if during application season you get a thick envelope and some bad days if you get a thin one. But happiness set point is a real thing. You will have good and bad days at either university, or job, or in your life, regardless. In fact, very close to the same number and intensity of good and bad days. Just like people who win the lottery or become paraplegic return to their own set points in mere months. If the devil appeared and said he was either taking away your ability to walk, or making you not get into your top choice for college, which would you pick? (And guess what – you would be okay either way.)


ADDENDUM: The Community College Strategy

People often attend community college with the goal of transferring to a four-year college, and ultimately, getting the same degree everyone else does. Community colleges accept all comers, and if the stressful part of this process is getting accepted to a four-year college, then isn’t this a very obvious back door? Yet I don’t hear many people discussing it. On one hand, I wonder if people avoid this seemingly very workable strategy due to the (stupid) status stigma of starting at a community college; on the other, I don’t know the numbers and maybe it is actually quite hard to transfer to a decent four-year college from a community college even for motivated people. There’s also the consideration of missing out on networking opportunities – some colleges are very cliquish, and landing there junior year when all the relationships are already established might be a very lonely experience, and you miss out on the friendships and career connections that are an important part of your career.

Bottom line, if you want to attend a good public university in a state, and you can't get in straight from high school, then if you're actually motivated and you attend a community college in that state, that's a realistic shot at getting in as a transfer student.

However it was hard to find statistics along the lines of "x % of people entering community colleges stating that they wanted to transfer to a selective four-year institution were able to do it." What I did find are some stats for specific universities (decent list here.) At UCLA, 30% of its undergrads are in-state community college transfers. Top public universities do take lots of community college students. (Interestingly, community college transfers outperform the direct-from-high-school students at top universities, but private schools don't take many - indeed, Princeton only recently started taking them at all again. Why might that be? One theory that explains observations is that Princeton isn't looking for performance throughout its student body, it's looking for alumni contributions that give them some mediocre legacy students, along with a "certain type" of applicant. Again, if you're the applicant, that's not about you or any choices you've made.)

You might also know that community colleges have an abysmal rate of its students ultimately graduating with a four year degree, so how does that match the higher success rate of community college students once they get to a four year campus? Because it's not the community college or four year college "making" the student - the community college students that succeeds at UCLA might not succeed first time around in the initial admission but they get to the same place, because of their innate properties. And if you go to community college and don't graduate - higher ed isn't for you, and that's fine too. Also worth noting - once you're at a UCLA, more selective private schools are within reach for graduate education. (But again - why? Ask yourself very seriously.)


FOOTNOTES

[1] If you're worried about a meaningless life: don't be. This is literally the thing that you have the most control over. As well, happiness is composed of three components, pleasure, meaning, and flow. With respect to these things, humans discount the future in two different ways. We value pleasure over meaning too much in the short term; that's easy to understand. But we also value meaning over pleasure too much in the long term. If you go years sacrificing food, fun, and friends to work like a dog at your meaningful job, you will likely burn out, and you may not get as much meaning from your career as you thought when you considered it for sixty seconds while you were 17.

[2] I am a physician, and did not get into the (highly competitive) top-ranked residency programs I ranked first. This was devastating at the time, especially because I realized some decidedly non-clinical-skill-reflecting poor choices likely played a part (i.e., it didn't occur to me to go over my application with my own med school, which includes getting a professional photo done, and I sent out my application with a headshot that made me look like a serial killer.) I complained to a friend that the people who did get those spots were just status-seekers who played the game better than me, to which she replied, "How is that not what's motivating you? And are you admitting that how well you play the admissions game is the important thing about your life?" Once I realized that I had become a physician to (news flash!) help patients, and that I would still be able to do this just the same, I got over myself. And the program I went through ended up giving me opportunities I never would have had at the other ones. Now, if I had gotten into the more competitive ones, would I not be writing a blog post about how that program ended up being the best one? Probably. Because it doesn’t matter to my overall happiness.

[3] People with narcissistic personality disorder have very little empathy or interest in others beyond using those others as a constant source of attention and flattery, so that the narcissist can convince himself he’s not garbage. Narcissists have very little self-value other than knowing where they are in status hierarchies, and very little knowledge of other people beyond categorizing them as higher or lower status. And narcissists as a rule are miserable.

[4] While on my surgery rotation, I met with the director of the clerkship, a trauma surgeon. He asked what specialty I wanted to go into. At the time I already knew it would either be psychiatry or neurology. But I was on my neurosurgery rotation at the time. In reality, for multiple reasons, there was never any danger of my becoming a surgeon. But the game I had been told to play was to include my current rotation in my list of interests in the hopes of avoiding offending the attendings and getting a better grade. So I did that – and the trauma surgeon looked contemplative and said, "I'm going to be honest with you. I don’t see you as a neurosurgeon." Of course he had seen through me and was exactly correct. There was a second of fear ("Oh no, he’s going to fail me") and then a feeling of immense relief – a physical sensation of weight lifting from my shoulders. That’s when I realized that my not being a neurosurgeon had been determined many years before that moment.

[5] Neurosurgery sounds cool but the training is especially grueling, even by doctors' standards. When I was a med student, I was in a seminar where a doc was going around the room asking what specialty people wanted to go into. One guy said neurosurgery, and the doc held out his hands as if weighing options and said "Yeah...(one hand) neurosurgery...(other hand) happy life...(one hand) neurosurgery...(other hand) happy life..." As it happens, I love what I do, psychiatry, which has a much more humane schedule than neurosurgery. I can't imagine why everyone doesn't do it, other than they don't think about what it would be like at age 50 to have a neurosurgeon's schedule.

Ranking of University Attended Does Not Correlate Well With Adult Well-Being

More on this in a Q&A format in the post right above.

It's worth understanding the relationships between the multiple statistics about a university,[1] and with its output - that is to say, life satisfaction. There are numbers for average early career salary, but precious little on satisfaction: Forbes, for example, uses a very dubious metric of percentage of alums donating to a university, and how much they donate on average. (Dubious, because this likely has to do with the institution's proactivity in soliciting donations, as much as with alumni gratitude.) There is almost nothing on overall life satisfaction - which is curious, because this is the main output we're concerned with when we apply, or send our kids, to universities - isn't it? What I could find, strongly suggests that the differences in life outcomes caused by universities are minimal, if any; the strongest is a ~10% premium in income at the most selective institutions. But the only reason to care about money is because we think it would make us happier - and if it's not, which is what the few numbers we have are showing, why are we bothering with this?

It's scary to buck trends, especially for parents, in a way that you worry might damage your kid's future. Here's what I'm telling you, parents: you're causing your kids definite harm with the current get-into-college rat-race, for benefits which are at best slight, and maybe - probably - are entirely illusory. That is to say, you may well be taking away a happy childhood for nothing. When you tell your kids not to cave to peer pressure, set an example for them to follow. If you're not brave for them, no one else will be. If you go all-out in college admissions madness from preschool on, you owe it to your kids to know why you're really putting them through this. For their sake, I hope it isn't just because you're trying to win status points or avoid judgment from your family or neighbors or coworkers.

The most important conclusion is for stressed out kids and their stressed out parents: you will be fine. You will get in somewhere, and you will get a good education, and have a good life, and it will be fine. The ranking numbers are often based on very arbitrary decisions, cardinal rankings are not good bases for statistics because they often imply gigantic differences, and the makeup of the individual students is far and away the most informative driver of choice of university. Kids: try to get into the best school you can for what you're interested in, but don't kill yourself to do it, and don't despair if it doesn't happen - because it really doesn't make much of a difference. Also consider where you want to live, and what kind of people you want to date and be friends with for life.


FOONOTES

[1] If you're interested in relationships between those statistics, they're below. I have to emphasize, again, for something that causes so much stress and consumes so much time, when we choose colleges, we really don't know what we're buying. There is amazingly little literature on outcomes, which suggests that whatever is driving the college admissions Olympics, it's not how much getting into good college benefits our lives. Even though that seems absurd, it's also obviously true.

For the relationships between SAT, acceptance, and endowment size, here's how it came out. SATs are more closely correlated with the other two, about equally with both in fact. Endowment per student correlates markedly less well with selectivity.

• SATs vs percent of applicants accepted: R^2 = 0.5896. The ones that are much more selective than their SATs would predict - the four military academies (which are obviously selecting on something besides SAT), followed by CUNY-Baruch and Babson College. The ones that benefit students - they are NOT as selective as SATs would predict - are Villanova, University of Maryland-College Park, and University of Minnesota-Twin Cities.
  
  
• Endowment per student vs SATs: R^2 = 0.5711. I even chose a logarithmic curve, since the SAT approaches but cannot pass 1600, but it didn't improve the goodness of fit.
  
  
• Endowment per student vs percent of applicants accepted: R^2 = 0.3388. The curve looked like endowment might have an increasingly marginal effect beyond $500,000/student, but even taking out the institutions above that or trying to fit exponentially or logarithmically didn't do much to the fit. The ones that are off-trend in a way that benefits students (lets higher amount of applicants in than their endowment would suggest) are Grinnell, Wellesley, University of Richmond, Texas Christian, and believe it or not Princeton. "Benefit" assumes that the endowment actually affects student experience.

[2] I had also been quite curious about the effect of nationally prominent athletic programs, especially football, on academic rankings. This is from personal experience, since I recall how the yield (% of accepted students actually matriculating) went up after Penn State's almost-number 1 1994 season. The Flutie Factor (cited in this paper by RT Baker, which was submitted to an academic institution but doesn't look like it's peer-reviewed) shows that the effect was known prior to that (the year after Doug Flutie won the Heisman, Boston College had a 25% increase in applications.) While there is very little literature about the effect of sports performance on academic ranking, this paper argues that in fact increasing football ranking does increase academic ranking.

Complex Dynamic Systems Like Cells, Humans, and Nations Cannot Avoid Cycles of Paroxysmal Disillusionment and Suffering

The following principles apply to any dynamic complex system, including organisms, individual people, or organizations - corporations, nations, or religions. They demonstrate that distortions will inevitably accumulate in the behavior of such a system, causing paroxysmal shocks and suffering.

Let's define a system as a set of components discrete from the rest of the universe. Let's define dynamic complex systems as ones that are high-information relative to the rest of the universe around them, high enough that in the absence of opposing forces, entropy passively favors irreversible changes happening to the arrangement of the system's elements which make it no longer discrete.[1] If such a set of components that exists now is going to exist more-or-less unchanged in the future, it has to perform actions[2] - it has to be dynamic - which reduce its own entropy at the expense of the rest of the universe.[3] The rest of the universe may include similar systems.[4] (Going forward, when I say "system" I mean "complex dynamic system.")

The seven principles are outlined below, then described in more detail.

1. A system must be mostly concerned with its own perpetuation, or it will not persist. "If a complex dynamic system has been around for a while, it's designed to last and expend energy to do so."
   
   
2. It must do this by reflecting - reacting to - aspects of the rest of the universe. Aspects of the rest of the universe important enough to make the system tend to develop ways to change its state in response to them are called stimuli. A more familiar statement of this principle: "That which gets measured, gets addressed."
   
   
3. A measurement always contains less information, and is therefore not a full or fully accurate representation, of the thing being measured. Perception cannot exactly be reality; "this is not a pipe."
   
   
4. Over time, the focus on self-perpetuation leads a system to become concerned with itself to the point of minimizing the importance of or responses to aversive stimuli to avoid altering its state (which is also aversive.) "Everything that gets measured, eventually gets gamed" or "All metrics eventually become useless." (Also known as Goodhart's Law or Campbell's Law.)
   
   
5. The system's responses become increasingly un-moored from the external world, favoring its own perpetuation over other functions, and/or having a severely distorted model of the world and reaction to stimuli. "Eventually, everything becomes a racket [and/or gets delusional]." Another term for this is representational entropy.
   
   
6. The distortions accumulate until a sufficiently destructive stimulus occurs (a shock), which either reorients the system (usually accompanied by severe aversive stimuli), or destroys the system. If the system survives, such shocks will happen repeatedly, but necessarily unpredictably. This is called inherent cyclic crisis.
   
   
7. This is inherent to any self-perpetuating dynamic complex system, and because these shocks are correctly perceived as worsening survival, they cause inevitable suffering. We can call this the Final Noble Truth, a vague parallel to the Buddhist First Noble Truth.

Following is an expansion on each principle.

Principle 1: "If a complex dynamic system has been around for a while, it's designed to last and expending energy to do so." If a system is going to continue existing, a top priority on self-preservation is mandatory, and self-preservation must be the primary influence on perception by and reactions of the system. Since Darwin, thinking of organisms in this way is not revolutionary. But the same principle applies to any other complex system, including human organizations. Corporations have a relatively clear function in this way (they can't keep making money for shareholders if they don't keep existing) but it's more surprising for most of us to think of religions, countries, or volunteer organizations in this way. The converse: if a complex dynamic system is not expending energy on its self-perpetuation, it will not exist for long. (Many apparent mysteries, like the transparently weak business plans of many a Web 1.0 company, are resolved on realizing that they are not in equilibrium and will perish quickly. And indeed they did.)

Principal 2: "What gets measured, gets addressed." This seems obvious enough, especially to those of us at all interested in organizational dynamics, but principal 1 directs the kinds of things that have to be measured if the system will persist, and principal 2 says that there is always some difference between the measurement, and the totality of the outside world; that is, there is always going to be at least some important information missed, and what is missed cannot be acted upon. Concretely speaking, genes reflect the outside world by establishing sensor networks that interact across the inflection point (the cell membrane, or in the case of multicellular organisms, the body.) Some sensor networks have become very rapid and fine-grained reflections of the outside world - xenobiotic metabolism enzymes (which have reacted in only millennia, and the genetics differ considerably even between groups of humans), adaptive immune systems (which also differ between groups of humans and react in minutes), and of course nervous systems are the paragon.[5] But all of them make sacrifices and do not (cannot) sample all of the possible information available. It should be pointed out that living things do not have to constantly repeat the mantra "what gets measured gets addressed" because that's how they're already built and behave, automatically and obligatorily, as was the case for eons' worth of their ancestors. This is not the case with human organizations, which are new developments in nature, and may not be in equilibrium - so those organizations that you notice do not measure important (survival-supporting) data are unlikely to exist for long either. Whatever corporations or their descendants exist in a million years, it won't be ones that didn't respond to relevant metrics. (Added later: A more generalized description of this process is called value capture, as described by philosophy professor C. Thi Nguyen at the University of Utah. Very briefly, value capture occurs when an individual encounters a group with a simplified, quantified version of their own values, often driven by considerations external to the agent. The additional component relative to "what gets measured gets addressed" is that in value capture, there is feedback from the simplified quantified value system to the individual, distorting and oversimplifying their own values.)

Principle 3: Perception cannot exactly be reality; "this is not a pipe." Representations - measurements - are never the same as the things themselves, and incentives are never aligned perfectly with desired outcomes (almost trivial; otherwise, perfect alignment means identity, that is, the incentives are the desired outcomes.) There must always be a limit on information collected, and inferences are not always correct. There is infinite information a system could in theory collect about the universe (looking for correlations between each datum or set of data), but the system is more likely to perpetuate itself the more the information it collects, and the more impactful the information it collects. It is this design choice by the system to sense survival-relevant data that turns one of infinite facts about the world around the system into a stimulus. Obviously, which things it chooses takes as stimuli - what it measures - are relevant. (Not to mention if the system is in competition, especially with others using the same resources, there are time and resource limitations on the system on how much data to collect before altering its state.) The implication is that there is a limited set of information collected out of all possible information - what the system receives as stimuli - and these stimuli are necessarily very heavily biased toward self-perpetuation.[6]

Principal 4: "Everything that gets measured, eventually gets gamed." This is similarly familiar, and here is where the tension is set up. Systems must perceive (measure) and react to their environment. Their measurements are not the same as the thing in the environment, only a reflection. Because of this, systems react to the measurement - the perception - not the thing that is being measured. This is not a trivial difference. Anyone who has worked at a large corporation or applied to professional school is familiar with this, and we all know examples where an endpoint was achieved in an only-technical, meaningless way that did not advance toward the real-world goal outside the organization that the endpoint was meant to incentivize. To "follow the letter but not the spirit" is an aphorism which expresses this. Case in point: many companies have sexual harassment or racial sensitivity training. These often take the form of instructional videos with quizzes after them. Most people skip and fast-forward through the videos as fast as possible to register as if they watched the whole thing, often having two different types of browsers open so when they get to the quiz they can go back if the answers aren't obvious. Of course this raises the question of whether there are some types of training where the written tests to get the credential have nothing to do with performing the actual work. For example, in the early-to-mid twentieth century, people became scientists because they liked being in labs, were good at organizing experiments, and in general got immediate feedback directly from their work, and therefore performed better, and therefore were recognized for it by peers and superiors, moreso than is the case now.[7] Do the best GRE scores (and administrative maneuvering, and recommendations, and tolerance of modern graduate school politics) really correlate with the best scientists? Or, does the same process produced skilled and caring physicians? (See Paul Graham's essay The Lesson to Unlearn.) For examples in individual humans, take your pick of any of a host of brain receptor-tricking molecules like opioids or alcohol, as well as immature psychological defenses like denial. Cancer is another example. Multicellular organisms have many checks on uncontrolled cell growth and spread, but there is a perverse natural selection our entire lives for cancer. Unlike infection or physical injury, cancer doesn't hurt until it's about to kill you, thanks to both the earlier reproduction of our ancestors relative to the later onset of cancer, along with the black swan of radiation and brand-new-to-nature chemicals that there was never any reason to develop an alarm system for.[5][8]

Another type of distortion has to do with the structure of the system, which effects the way it behaves, rather than perceptions per se. It's long been noticed that corporations become less "nimble" (responsive to market change; ie the relevant universe outside the corporation) as their surface area-to-volume ratio goes down.[9] The higher the surface area-to-volume ratio, the more information that can be collected and the more effective responses can be. Think of bloated giants like big automakers or old engineering companies, where in Dilbert-like fashion people think more about maneuvering in their jobs, coordinating with other departments within the company ("transfer pricing") or competing with other people within the company than they do about their outside-the-company competitors or serving the market. This certainly occurs within states as well, where to various extents for Chinese dynasties, the Roman Empire, and the Spanish colonial empire their downfall was more the result of special interest maneuvering and other intrigues directing attention consistent inward to the court - because what could be going on that's more important outside the palace walls where the barbarians are than inside where the power is? So we have Zheng He's fleet being recalled, Roman patricians scheming in the absence of a succession rule to get legions on your side, or Spanish royalty neglecting overseas possessions until the British and their offspring eat your lunch.[10] There's a final type of distortion which arises from the way that nervous systems save time and money: the more some stimulus-response pairing occurs, the less reward-sensitive it becomes. It moves from being a conscious act requiring effort and concentration, to a habit, to (in biology, programmed through evolution) a stereotyped movement, then a reflex. Once a stimulus-response pairing has started moving down the line it is almost impossible to move it back other than by over-writing it with another stimulus-response pairing.

Principle 5: "Eventually, everything becomes a racket [and/or gets delusional]." This quote is attributed to the late George Carlin. An occult paraphrasing of this in more specifically politics terms is "The state is primarily concerned with itself." Here we can see the full flowering of the problems buried in the earlier principles. There is a constant tension between the need for negative feedback, which the system avoids - that's what aversive stimuli are for - but because the measurement and the thing measured are not the same, the metric is game-able, and the system avoids these stimuli increasingly by gaming them rather than taking real action. (And simultaneously makes un-helpful-to-survival end-runs to pursue positive stimuli.) Consequently the stimulus-response arc gets more and more distorted with respect to the actual longer-term perpetuation of the system. This seems paradoxical in light of #1 above, but because systems are never perfect reflections of the universe around them, they necessarily always react based on at least limited information (especially with respect to long-term consequences) and sometimes with outright distorted information their machinery is feeding them. The necessary self-focus means that these distortions will tend to be in favor of pursuing pleasure; of avoiding pain and believing everything is alright when it is not, and over arbitrary time whatever non-self-perpetuating parts of a system's "purpose" previously existed, will atrophy, and its behavior will become more distorted in favor of comfort and perceived survival over actual survival.[11]

The distortions come not just from "gaming" pain. Organisms can hack themselves to fire off their reward centers without an actual survival enhancement, for example, with heroin, masturbation, or consuming resources they are driven by prior scarcity to consume as frequently as possible, but which have become "cheap" to the point where their over-consumption causes problems, e.g. in humans, fat, salt, and sugar. Opioids are in humans the thing closest to the artificial intelligence problem of "wireheading" where a self-modifying agent given a task can self-modify to be satisfied even though the task is completed.[12] Good examples of rackets are religions and charities that depart from their stated mission in favor of wealth-accumulation and self-perpetuation. (See Givewell's list of charities which maximize their mission rather than their income.) Profit-seeking entities whose products or services intrude into "sacred" (i.e. non-transactional) realms (best example: healthcare) often find that self-perpetuation wins out over their claimed mission. Organizations and individuals can also become delusional - humans are incorrigibly overoptimistic and discount the future.

Principle 6: inherent cyclic crisis. Eventually the stimulus-response arc becomes so distorted that it encounters a survival-threatening problem it can properly perceive and respond to, but by this time the gap between perception and reality is profound and it comes as a shock. Surviving the crisis, if possible at all, is quite painful.

Black swans are indeed one type of crisis, but missing impending black swans is the fault of the system only to the extent that the system could not reasonably have anticipated the black swan event, given the experience it had to draw on. More salient here are crises precipitated by the accumulated distortions in the system's perceptual machinery, where the system "should have known better". At the organizational level, nations might collapse because their ideology, increasingly un-moored from reality, led them to weakness on the battlefield out of refusal to update their armies with modern techniques and technology. Nations with dysfunctional (delusional) organization meet reality catastrophically on battlefields, and religions collapse (sometimes) when encountering reality. Crusades failed due to Christians' belief that God would intervene, medieval Europeans with a military hierarchy based on nobility got crushed by Mongols with rank based on meritocracy, Washington in the Seven Years War lost to the French because he insisted on fighting like a gentleman in rows, then the British lost to Washington in the American Revolution because they still insisted on this formation, and Washington no longer did. (Many of these could be considered examples of the advance of "rational" (and more destructive) warfare over traditional warfare.) For the young Washington and the later British Empire, the losses did not destroy them but came as painful shocks. In the case of many near-delusional Crusaders or the combined German-Polish-Hungarian forces in thirteenth century Europe, the shock did result in their destruction. On the individual level, any delusional or distorted behavior (psychosis, neurotic defenses, substance use) results in a painful shock in the result of adjusted behavior or shattered beliefs, or in some cases, the death of the individual. Someone might underestimate the risk of driving while intoxicated or in inclement weather, and crash, injuring or kill themselves or their family and updating their belief only in crisis. These crises occur more often and faster, the less (or more distorted) the feedback, as illustrated in very centralized arguments from authority (famine under Stalin using very divorced-from-reality - and unquestionable in Stalin's USSR - Lysenkoist theories of biology and agriculture.)

Principle 7: What to do about it? The Last Noble Truth is that cyclic crises are therefore inevitable in any complex dynamic system. As conscious complex dynamic systems called human beings, composed of complex dynamic systems called cells, being members in conglomerations of complex dynamic systems called nations and corporations and politics or religious belief systems, we will occasionally have shocks that kill us, or even when we "wake up" and adapt, still hurt us quite a bit. This happens in national collapses and revolutions as well. In arbitrary time, the problem will always re-emerge. Your measurement is not the same entity as the thing it measures. Unless a system comprises the whole and then there is no boundary, and of course no system.[13] How can we minimize the inherent problems that lead to this cycle?

• Constant testing and cross-checking between senses and expectations. In individuals we already do this automatically (corollary discharge, binding between senses and discomfort when our binding expectations do not match observation.) Cross-checking beliefs and assaying decisions at multiple points in ways that will quickly expose them if they were bad is helpful for individuals; it probably won't hurt to think twice about that turn you just made while hiking in the deep dark woods, even if you feel quite confident about it. Critical thinking is one form of this. Cognitive behavior therapy is another.
  
  
• Increasing the amount of feedback. This facilitates the suggestion above as well. It is good to decrease the consequence for objections in groups. Calibrate yourself - when people or organizations have secure egos and want to get better at something - running a mile, or making better decisions - they do this. This experiment about how to fool a computer in a "delusion box" showed that through a constant drive for being surprised - by learning new information - an agent gets out of being deceived faster. Of course this itself is also hackable (the machine could reset itself; you could convince yourself you're learning important new information but really you're just confirming your biases.) Pushing until you reach failure, in physical training or decision-making, is an instrumental rather than epistemic form of increasing your surprise.
  
  
• To the extent possible, rely on positive feedback. Negative feedback is that which by definition systems avoid, and they will avoid it by gaming it if necessary. Therefore, systems should put themselves in situations where the ratio of positive to negative feedback is higher, so we are less likely to avoid feedback.
  
  
• Simulating negative outcomes. In other words, expect the worst. You will never be disappointed, and you will have prepared yourself for the shock. Negative visualization as proposed in stoicism is a technique to do exactly this.
  
  
• Respecting a system's realistically unmodifiable constraints - especially if the system is you. This is especially true at the individual level. Humans as a species are not built to question close family relationships, especially without reason, without becoming depressed and damaging our relationships (asking if your daughter really loves you once a day will not help you or her.) For that matter, negative visualization actually causes quite a few people to reliably suffer rather than feel better (including me.) Constantly second-guessing every decision, like that turn you made in the woods, may erode your confidence and spark depression. Of course, your daughter really might secretly not love you, and your family might die, and you might have made a wrong turn (and you really can't fail without consequence at some things.) But it's probably going to make you suffer more in the long run not to think about this all the time, and you should pay attention to your reactions to see where your thresholds are.

FOOTNOTES

[1] Though not the purpose of this argument, this does set up a useful boundary for defining living and non-living things. Because it's quantitative it doesn't suffer from a problem of boundary cases, and sentience is nowhere considered, but it does appeal to common sense - stars and fire are at one end of the spectrum and things with nervous systems are at the other. Most current electronics are minimally dynamic, which makes them "less alive" than is often argued.

[2] It's implied that these actions are cyclic, like catalysis in biology or the Krebs cycle, or else the system would be unlikely to return to any previous state, and you don't have a self-perpetuating system.

[3] There's a strong argument to be made that defining a system as separate from the rest of the universe is arbitrary. However this becomes less true as the system develops additional complexity for its self-preservation. There is an increasingly sharp inflection point at some physical boundary of the system where the exchange of matter and energy between unit volumes drops, and also where an event on one side of that boundary events have much more impact to the future of the system than events outside of it. That is the self/non-self boundary. In cells this is easily recognized as the cell membrane. In nations, although the boundary becomes more complicated, ultimately the boundary is spatial, because of the primacy of space. Even in corporations or religions this remains the case. The individuals in those organizations, or carrying those beliefs, as physical beings are still dependent on predictability, safety, and resources permitted by more "basic" forms of organization. A good example is the early evolution of life, it is recognized that an RNA molecule (or RNA-protein) would not benefit from any reaction it could catalyze any more than another molecule in its vicinity, or at least not as much as it could if the reaction products were sequestered. Consequently when nucleic acids were enveloped in lipid membranes, natural selection accelerated, and the the self/non-self boundary became less arbitrary.

[4] In a zero-sum setting with limiting resources (which is a necessary condition given arbitrary time) this is a good definition for competition. Unless you count Boltzmann brains, it is likely that a system will find itself in a world with other similar systems.

[5] You may have noticed that there are no examples "below" humans in my examples in principle 4. There are many examples of behaviors in humans, and in human organizations, where metrics are gamed. There are far fewer examples of organisms besides humans where this is the case. Some species of pinnipeds dive deeper than any prey we know of, we think, just to alter their consciousness, and African elephants go out of their way to consume fermented (alcoholic) marula fruit. But there is nothing like the systematic distortions we see in human psychology. It seems likely that the simpler, the more fecund, the faster cycling an organism is, the less it is able to afford having a gap between its response to its metrics and what survival-affecting things are actually occurring in the world. That we don't see many organisms gaming their metrics could occur both because their stimulus-response arcs are simpler, and because a distortion in these arcs will more quickly kill off the organism so they don't come to our attention. This also implies that having a mind as complex and powerful as ours provides unique opportunities for distortions - that organisms which are focused on "survival, unclouded by conscience, remorse, or delusions of morality" (to quote a psychopath fictional android) in fact have a more accurate view of the universe (not incidentally, a central theme of the Alien films.) Having a complicated brain and surviving by imitating each other, our nervous system is constantly hijacked by self-reproducing ideas in a way that our genomes never were (lateral transfer events are incredibly rare) and those memes are selecting, as per principal 1, for their own self-perpetuation; they want to avoid killing us outright, and use us to spread them, so they can be slightly parasitic, occasionally symbiotic, but most are probably merely commensal. If there is an analogy, the meme complexes we get from our families are not genes but rather our microbiome. It should also be pointed out that as we congratulate ourselves for taking over the planet due to agriculture and combustion engines, we are living through the sixth mass extinction, suggesting that in fact we are not acting in our long-term best interests; and like cancer, ecocide might not hurt enough until it's too late, and that intelligence is an evolutionary dead end (a principle 8? That given arbitrary time, an inherent cyclic crisis will eventually destroy an entity, and the more complex the representational system, the more likely the crisis is to be endogenous.)

[6] There is a spectrum of arbitrariness, of how "symbolic" the perceived stimulus is relative to the thing being perceived. In the engineering of signal systems, the closer your signal is in a physical causal chain to the stimulus - the thing it is signaling about, or measuring - the less arbitrary it is. Digital systems are more powerful in many ways than analog systems but they accept increased arbitrariness and complexity in exchange for this. Case in point: people who fear assassination can build elaborate electronic sensing systems to avoid being approached while they sleep, but there are always questions: can they be hacked? What if a component fails? What if the power is out? Can a spy shut it down? Compare this to the system used by the Tokugawa shoguns - sleeping in the middle of of a large room with a wooden squeaky floor with many tatami barriers, and choosing a place to sleep on that floor at random each night.

[7] Of course other things have changed about the way science (and medicine) are practiced over the past half-century, not to mention that all the "low-hanging fruit" in terms of problems accessible to the specific strengths of human cognition may have been picked soon after the Enlightenment started. But it remains a concern that by (not unreasonably) trying to regularize and make transparent the application and career progression process, we're selecting for attributes that have little to do with being a successful scientist or physician, or even selecting against them, because we're using "artificial" endpoints distant from the relevant abilities, which can be and are gamed. Certainly this problem is not unique only to science and medicine, and whatever is causing the phenomenon, it's having real-world economic consequences. An interesting historical study would be to see if the health of the Chinese government across the dynasties waxed and waned with any relation to some aspect of civil service examinations.

[8] Referred visceral pain is an example of an aversive stimuli-sensing system that gives very inexact answers, because it was never important enough to improve. If your arm is hurt, you can point to exactly where even with your eyes closed. But when people get appendicitis, very often in the early phase they point to the center of their abdomen around their belly button, and then gradually the pain moves to the area immediately over the appendix - but only after the overlying tissue, which is innvervated by somatic ("outside"-type nerves) is irritated. Often people with a problem in their abdominal organs or even their heart feel extremely sick and anxious and in general uncomfortable but can't point to any specific spot. Why does this make sense? If a scorpion scrambles up onto your left elbow is stinging you there, it's worth knowing exactly where the stinging is happening so you can act in a way that improves the situation. But if you were sitting around a fire with your tribe in the African Rift Valley 100,000 years ago with appendicitis, what exactly could you do about it? If you had bad stomach pain, it didn't matter exactly where it was; you curled up in a ball where your family members were nearby to care for and protect you and hoped it passed.

[9] In contrast to corporations, single-celled organisms survive best not when they have a low surface-area-to-volume ratio (like successful nimble corporations) but a low one, which is why they are mostly near-spherical. Corporations, while competing with each other and in some ways with their customers, are still operating in an environment that is predominantly cooperative, so it's better to have lots of customer interaction surface. Bacteria exist in an environment of constant unpredictable ruthless lawless natural selection. It's really about how much the surface is an asset for information gathering by the system, versus a liability to attack from competing systems. Consequently, for bacteria, the sacrifice of knowing less about the outside world (which at that scale is less predictable than our world anyway) must be worth it given the overall survival advantage gained by being in the shape that most maximizes distance of any unit volume from the surface. In contrast, there are cells in biology that maximize surface: neurons, and nutrient absorption membranes deep in the GI tract. Both of these exist deep in the organism (especially neurons) in a web of profound cooperation (also especially neurons.) Of course, some bacteria are rod-shaped, which is a higher surface area-to-volume shape thought to confer an advantage in terms of nutrient absorption - when that is the growth-limiting factor rather than attack by other systems, that is the shape we would expect to see. The more fractal a complex dynamic system, the more likely it is to exist in an environment of predominant cooperation. The more spherical a complex dynamic system, the more likely it is to exist in an environment of predominant competition. In the case of corporations, the shape is somewhat "virtual", but corresponds to points of contact per customer and ease of contact, which ultimately are still going to require space. Nations are somewhere in the middle, though it would be interesting to see if nations now, more cooperative and less violent than they historically were, are more likely to have fractal borders, or shared zones (my predictions) than one or two centuries ago.

[10]This corresponds to Level 3 operations as described here, which explain how large organizations work but is not an argument that they should work that way, for Level 3 organizational decisions often lead to the downfall of the organization, unless the inner circle in the super-Pareto distribution has the best interests of the organization at heart. In politics, maintaining the state's best interests happens either because the leaders have a feedback loop in the form of being beholden to an informed electorate as in functioning democracies, or less likely, by luck as with benevolent dictators, e.g. Lee Kuan Yew of Singapore, or the "good" Roman Emperors.

[11] There's an apparent conflict here. On one hand I'm arguing that systems become distorted because they're focused on self-perpetuation and ignore metrics. On the other I'm arguing that they focus on the metrics, which they game and become distorted. Both can be true at different times, and vary depending on the aversiveness of different stimuli.

[12] Organisms are exempt from becoming "rackets." Rackets are systems which have a claimed mission besides their self-perpetuation but in fact are only self-perpetuating, and organisms are openly survival systems, full stop, and make no claims to the contrary. In other words, organisms cannot be rackets, because organisms do not have mission statements to deviate from.

[13]It may not have escaped your notice that one implied solution - expand the system until it comprises the whole universe, and there is no self/non-self boundary - is, at least on the individual level, one advocated by many mystical traditions. We actually achieve this when we die, so in individual terms this could be re-formulated as "lose your fear of death". Yet our read-only hardware makes this a terrifying and unpleasant experience, even, empirically speaking, for life-long meditators. For now, this is not a real solution.


APPENDIX: Analogous Terms

	Non-human Organism	Human	Organization
Input	stimulus	perception, belief, representation	measurement, metric, dogma
Reflection, Output, Reaction	response	belief, behavior	decision
Examples of Gaming	Rare; some higher animals seek out "highs"	Opioids, denial, delusion	Preserving letter but not spirit; false or incoherent religious or political truth claims
Examples of crisis	Death	Personal disillusionment or death	Revolution or collapse