I wonder, to some extent, how much things like this are reflective of the ascendancy of an intellectual movement that had some real and rational critiques of what came before it but has more or less failed to establish a positive program of its own.
My wife has a masters in urban planning. The acquisition of said masters involved a great deal of, basically, self-flagellation on behalf of the discipline of urban planning, in which the students seemed to learn a lot more about the (genuine) ways in which urban planning royally fucked up the mid-20th century than about what it could do right using what we'd learned from that experience. (I exaggerate somewhat, but, like, not all that much.)
None of this was that shocking to me, because I'd gotten into Jim Scott in college in the 2000s and pushed him on friends like a cool new drug. High modernism wasn't exactly riding high in 2007 (although I guess the US government was actively attempting to summon a functional democratic capitalist Iraq out of military force and ultra-crystallized McKinsey so maybe I'm wrong) but nonetheless the critique was fresh enough that it grabbed our attention.
And so, you'd think, maybe out of this will come some new synthesized perspective, but it doesn't really seem like that happened. I'm a data guy now and I constantly emphasize the importance of being aware of what the metrics actually measure and steering clear of optimizing for misguided targets, so I've clearly been influenced, but the case you're presenting above sure sounds like a description of a critical movement that spiralled of joyously criticizing its defeated bogeymen rather than engaging with people who have internalized a few lessons from either direction.
I'm not averse to the argument but citing climate change as an example? Climate change is the poster child for technocratic over-confidence, wildly wrong predictions and the failure to balance costs with benefits.
Scott, I love your reasoning and think you have a great sense of which facts are important to a point, but when the facts *don't* matter, you almost invariably get them wrong. SCOTUS have JDs (or in the past LLBs) and Brown principally concerned the 14th Amendment, which was less than 90 years old at the time.
1) I'm confused about 1-5 - of these 1 and 4 seem objectively correct with later (scientific) knowledge, 2 seems to be correct in the sense that it's a no-brainer under our current values, 3 and 5 seem like a good examples of the state messing things up (lockdowns aside, the state *also* said "no masks" there for a long time).
2) A much bigger example of something that bothered me in Seeing Like a State - he tears up a lot of western agricultural practices that failed to generalize, but wasn't the green revolution a thing that happened, largely successfully, mostly eradicating third world hunger? What was different about it from the failure cases he's quoting? And most importantly, why does Scott never ask this? It reminds me of the three number pattern game in HPMOR - Scott assumes he has a theory to explain the failure of some big projects, but he never tests his theory on any government successes to see if it can separate them.
I was going to dismiss Weyl out of hand, as I do for anyone who uses phrases like "strongly privileges rationalist approaches". But I realize that I can fully consider what he's saying and still dismiss him, based on the axis I think really matters: Chosen Vs. Coerced.
But I want to first take a whack at a similar issue that I think Scott elides in his introduction: "Nobody ever defends technocracy. It's like "elitism" or "statism". There is no Statist Party. Nobody holds rallies demanding more statism. There is no Citizens for Statism Facebook page with thousands of likes and followers."
Consider for instance some comfortable English professor defending Russian totalitarianism. He cannot say outright, ‘I believe in killing off your opponents when you can get good results by doing so’. Probably, therefore, he will say something like this:
‘While freely conceding that the Soviet regime exhibits certain features which the humanitarian may be inclined to deplore, we must, I think, agree that a certain curtailment of the right to political opposition is an unavoidable concomitant of transitional periods, and that the rigors which the Russian people have been called upon to undergo have been amply justified in the sphere of concrete achievement.’"
For the record, I can give arguments against all 5 examples Scott mentions, and, yes, I can drag in quotes from black activists who said that forced integration was a fiasco, but if I explain what's wrong with all of the different initiatives we'll be here all day.
The point is that euphemism is always a way to evade what one is talking about.
Anyway, back to Weyl - the contradiction here is that he is against the E.A. movement which is purely voluntary. No one is forcing anyone to take part in it, and - as I understand it - everyone can make the case and present the evidence for and against it, and for and against what causes the E.A. movement tackles.
If you read the article, you see he somehow conflates both Stalinist central planning and economic liberalization. So the whole thing is a con job - it's saying: "Some people who thought they knew better advocated for _this_, and other people who think they know better advocated for _that_, so they're the same, and you shouldn't listen to either - because I know better!"
I don't think that technocracy vs "unnamed good way of making decisions" is necessarily a good or useful dichotomy, but I do think that there is value in looking at what various solutions or ways of doing things are optimizing for. For instance, school admissions. Though this seems a little like a methodology debate - which I think is what talks about technocracy really are, discussions on the best way to do things - it seems more deeply like a debate about what you want in a student body. If you are optimizing for a student body who can expose each other to a diversity of backgrounds and life experiences, your admissions process is going to look different than if you are just attempting to get the most intelligent students.
I think a real postmodernist or critical theorist or whatever you want to call them would say that power is what is missing from both your and from Weyls analysis though. Just because something isn't data driven or centrally planned doesn't mean it's less likely to perpetuate power differentials or intergenerational wealth or elite capture, it's just that "technocracy" gives things a veneer of objectivity that would otherwise be missing, which gives it more staying power than otherwise. At least, that's what I think of as the actual critique of technocracy - not that it necessarily has worse outcomes than other methods but that it pretends to be above the entrenched interests etc.
It’s very easy to agree with your essay in general, but counting the handling of COVID as a triumph for technocracy [outside of China] undermines the fundamental premise. The mask fiasco, the absurd discussions surrounding border closures, the idiotic conservatism surrounding vaccine approval, much of the school debate, etc were mostly elite failures. A good dose of Folksy Common Sense wouldnt have been terrible in many of those debates.
I don't know what this is a product of, but I do not have negative connotations for the word 'technocrat'. The main usage I associate it with is when, say, The Economist says Country X has appointed a technocrat as Minister of Salad Forks or whatever, i.e. someone chosen for their competence rather than affiliation to a particular party; it's invariably presented as a good thing.
Does the word have universal negative associations in other media circles, or just in particular ones that Weyl would be a part of?
"it worries me that everyone analyzes the exact same three examples of the failures of top-down planning: Soviet collective farms, Brasilia, and Robert Moses."
Excuse me, but Robert Moses is an example of a SUCCESS of top-down planning!
N.B. Since it appears to be impossible to find replies after I've been informed of them by email (or even to find my original comment), I'm going to put off responding to such until the site is fixed. In the meantime, I'll just sit here and pontificate.
Technocrats were more useful when 1) the world had much less information and 2) that information was only available to technocrats pre-Internet.
Today it's impossible for technocrats to absorb even a sliver of the world's information and they're often not much smarter than a hobbyist or blogger somewhere.
Technocrats also tend to impose highly centralized decision-making and solutions, when decentralized systems are more robust and experimental.
The really strange thing about the Interstate is that it was very likely heavily inspired by a nightmarish journey across the continent that Dwight D Eisenhower made when he was a young Lieutenant Colonel - one that must have made him think "There has to be a better way to do this!" - and it's hard to get more common-man and less elitist than "Army man made big remembers his American road trip"!
(Granted, during World War II Eisenhower was a general who had never seen combat, so he may be a disgraceful example of an elite holding positions of unwarranted high rank; but he did fairly well for himself despite his lack of experience.)
The only thing I can really think of here that maybe can be added is the idea of technocracy as the bureaucratization of government. Notably, I don't think their existence is a harmful features of technocracy, but rather that the process vs judgement is associated with principle-agent harms.
Which one I think is harmful depends on whether we're considering internal decision making or external relations. Internally, we see that having democratic systems of judgement (ie. commitees decide FDA approval vs simple rules or housing by right vs housing permit systems) adds complexity and delay, as well as undermining good outcomes.
I think the flip side is that having process based external relations vs democratic based external relations undermines responsiveness, especially if combined with internal judgement. It means that the personnel is unresponsive to outside feedback loops, which makes it hard for voters to ensure their best interests are being taken care of (eg vs. military contractors in the military or working towards liberal goals under R admin). The combination that I see as 'bad technocracy' is internal judgement (as opposed to formal process), which makes policy illegible for lawmakers, alongside external process (vs democracy) that makes it hard to effect personnel.
I don't think it's an accurate characterization to describe Weyl as taking the "judgment" side of the "mechanism vs judgment" axis! After all, he is a strong advocate for quadratic funding, Harberger taxes (aka COST aka SALSA...), superlinear quantity subsidies and other "radical mechanism design ideas". That said, he's certainly not anti-judgement in any sense either; I view him as being in the "we need the mechanisms, and we also need stronger public discourse to help us better use the mechanisms" camp.
I think the best way to view his position on this issue is that he has a strong opinion on the tradeoffs between different *types* of mechanisms. This piece co-written by him and myself is IMO a good explainer of the core idea:
Basically the idea is that mechanism designers should explicitly have simplicity and explainability as core values to a much greater degree than (math/econ) academics are naturally inclined to accept today. A more concrete example of what he's getting at is this piece from him:
That piece talks about his experiences in spectrum auction design, and how a very complicated auctioning style actually was abused and led to a few actors getting spectrum licenses at very unfairly cheap prices. Furthermore, the "illegibility" of the mechanism makes it inherently hard to prove to the public that something wrong happened, because defenders of the outcome can just spout some technobabble that goes way over the heads of not just the average person but even the smartest person that the average person listens to.
Both markets + property rights, and democratic voting, are far superior on this dimension. The mechanisms are simple enough that even lay people can clearly understand how the mechanisms are supposed to work and so they trust them. (Yes, I know, lots of people express suspicion about both property rights and voting all the time. But I'm sure you can agree that if people's cars and houses and who the president is were instead allocated using "theoretically optimal" VCG auctions, then things would be 1000x worse).
Furthermore, markets and voting are both not *opposed* to "judgement"; it's pretty clear that in both cases, they perform reasonably well precisely because a lot of complicated and (necessarily) illegible judgements are being made by the participants in the mechanism! Though clearly this is not true of all mechanisms; Twitter likes/retweets, for example, have no way to express concepts like "I think this tweet is socially harmful and should be shown to fewer people, though it's not quite so egregious that it should be removed outright".
I do disagree with parts of his EA critique. Particularly, it's worth noting that one of the projects more prominently favored by the EA community, GiveDirectly, involves *just giving poor people money and trusting them to do what they want with it* - in many ways the ultimate anti-technocratic form of charity. (Though EA does do its fair share of more-centrally-planned things, like handing out anti-malaria bednets, as well).
I think there might be a mistake in: "Some people analyzed a bunch of data, came up with an algorithm for diagnosing heart attacks, and told doctors they should use the algorithm instead of their own judgment. (...) The interesting part (search "Goldberg Rule" in that link) is what happens when you give doctors the algorithm and tell them to use it to supplement their judgment." - I just looked up "Goldberg rule" in the article linked and it seems to be about diagnosing psychosis, not heart attacks.
There are multiple Statist parties, two of which spend billions of dollars evangelizing each election season.
Also, during my undergraduate years, I once came across a bound copy of several issues of the Technocracy Party newsletter when I was supposed to be studying. They were a Vietnam-era group which seemed to be pushing for evidence-based socialism to be applied by scientific-minded leaders - probably would have been big Yang fans.
'Suppose you're the FDA trying to regulate cigarettes, and your doctors and bureaucrats have come up with a plan. What do you do now to "seek democratic feedback"? Put a suggestion box outside the FDA office?
Here in the UK - yes. There's a requirement to have a public consultation on new policy - anyone can answer. It's not a vote but you have to consider the arguments raised and if you just ignore things raised without addressing them you can get into legal trouble.
Strangely, Scott misses the opportunity to discuss the large scientific literature that actually tests whether "mechanism" predictions (and thus decisions) are better than "judgments" predictions, or what researchers call algorithms, or actuarial, or model predictions vs. human, clinical predictions. Fortunately, I wrote a blogpost summarizing this evidence a few years ago (2016), so readers can just read that.
TL;DR formal methods essentially always beat human judgments, even when given the same data, even when humans are given the predictions from the formal model. I interpret this in line with human's desire to be in control, which is apparently stronger than their desire for good outcomes. In these cases, it seems wiser to set strong defaults in favor of whatever the models are saying, and let some minority of humans opt out of that if it's not too much of a problem (maybe for a price).
I think more discussions on policy should draw attention to the fact that policy disagreements often hinge on different preferences (disagreements on facts seem to spring from a desire to support one's preferences). In this light, there is obviously no such thing as objectively correct policy, unless everyone already agrees on the desired outcome.
The issue with technocrats is that they often seem to think everyone agrees with their own policy goals, so all that's left is convincing everyone that their proposal for achieving this goal is optimal.
It’s interesting that one of the primary benefits of systematic technical decisions is legibility, when many systems being implemented by the state are increasingly illegible (either proprietary, or a result of inscrutable ML models that even the creators can’t explain). I wonder if Weyl would agree that technical decision makers actually *are* more legible. (Not disagreeing with the general point that scrutable >> inscrutable, just pointing out how that classification is getting less-broadly-true over time)
Others have pointed to this, but I want to present specific arguments as to why 4 out of the 5 case studies at the beginning aren't particularly strong cases.
1) Vaccines - I will skip this because I agree it's a good example of productive "technocracy", and it's the strongest of the 5 case studies.
2) Desegregation - The high court mandating an end to segregation was good. But the technocratic court-ordered implementation left much to be desired. By the 1970s, busing had an under 10% approval rate among all races. Just about everyone thought - and I believe most would still agree today - that as technocratic federal courts further embroiled themselves in the day-to-day planning of desegregation, without respect for local considerations, the worse matters became.
3) Interstates - The case study cites how mob/activist input caused highway plans to be changed from technocratically-determined direct routes to ones that caused less turmoil. Is the argument that these changes were bad? In my eyes, this case study shows mob/activist input causing a potentially-brutal technocratic policy to be changed to something more humane. So I don't think this is technocratic success; it's at best evidence of the success of a mixed approach. (Also, I think your discussion of it thus far doesn't sufficiently engage with the real pain that can come with the government deciding to take and destroy your home.)
4) Climate Change - I'll give you cap-and-trade, but that's not the most significant technocratic anti-AGCC policy at all. Most technocratic proposals against global warming have been a wash at best, or actively harmful. Natural gas is better than the status quo, and its rise is one of the biggest reasons US emissions are decreasing, but climate-conscious technocrats want to ban fracking because "fossil fuels bad". Technocratic consensus is for the Paris accords, despite simple back-of-the-envelope calculations demonstrating that they impose *nowhere near* strong enough burdens on e.g. China and India to have any significant impact. And here in California, climate-conscious technocrats chose gasoline taxes as the ballot issue to champion, despite most economists' recognition that gas is an inelastic good and this will not significantly lower emissions. Maybe certain smart people on Twitter have the solutions all figured out. But the technocrats aren't listening to them, and what they're actually doing in the present day is not helping.
5) Lockdowns - I don't want to dwell on this because I'm sure others will. I'll just say - literally today, it was made clear that the state with the highest death rate (and death count) in the country got there by intentionally ordering COVID patients to be placed in nursing homes, among the most vulnerable possible populations, and that they then covered it up. I know you probably wrote this before the news broke, but still, I think it might just give the lie to the idea that the technocratic response to COVID was a triumph.
A really good, nuanced critique of Weyl's essay, which quite annoyed me for exactly the reasons you describe above. Weyl in general seems to have gone backward over the last few years - I really enjoyed his book Radical Markets, but he recently withdrew support for some of the policies he recommended on grounds that to me seemed shaky and ideologically motivated.
Reading the comments, I'm noticing a lot of people conflating technocracy with centralisation; while this has been the case historically I don't see it as a general rule. Decentralised systems can absolutely be technocratic if they arise on the basis of rational and scientific principles. Bitcoin is a great example - the poster child of decentralised finance, based on a mathematical algorithm. In general, I'm very bullish about the application of the findings of complex systems science to governance. I think using principles listened from systems biology/sociology will allow technocrats to design systems that benefit everyone while avoiding the pitfalls of the old school high modernism. It needs a name - Systems Modernism, perhaps?
> "humanities, Continental philosophy, and the humanistic social sciences" - isn't that usually code for stuff like queer theory, postcolonial theory, and postmodernism?
"code for" seems like a cheap and unnecessary jab... I would agree, however, that the humanistic social sciences, etc., etc., *consist mostly of* stuff like postmodernism, etc., etc.
I'm not a New Yorker, but it seems like Robert Moses got a lot of big things done that make NYC more livable for its immense population. I'm from the San Fernando Valley, which used to do big infrastructure projects like turning the LA River into a giant concrete ditch after the 1938 flood and building the Hollywood Freeway in 1955. But very little has been done since, despite a big increase in the number of households.
My guess is that Moses' early infrastructure projects had high returns, but after awhile they had diminishing marginal returns while Moses became more set in his ways as he aged.
Thus, the huge backlash against him.
But I suspect Moses's reputation will rebound after Robert Caro is gone.
"Last time anyone checked, Gates had an approval rating of 76%, the highest of any figure asked about and literally higher than God."
This is slightly misleading; while Gates does have a higher *approval* rating than God (who had 52%), Gates' disapproval rating was 20% whereas Gods' was only 9%.
The entire premise of this post is wrong. Technocracy is not a Russell Conjugation, although it can be used as such. Technocracy is not evidence-based decision making, it is the creation of a veneer of complexity to avoid criticism and scrutiny of decisions, where only people who agree with the ultimate decision are allowed criticise the justifications for said decision. Its the facade of public scrutiny where none exists, because the people allowed to criticise will not upset the apple cart in the knowledge that they will lose their voice (and eventually their job security) by doing so.
Technocrats are not pilots taking us to new destinations is complex feets of engineering, they are scientistic frauds.
The incredibly basic differential equations and thumb in the air assumptions upon which they are based, calibrated early and unmodulated by subsequent reality, that were used to justify indefinite lockdowns are a clear example of this.
But even the smallpox vaccination protests may have been legitimate at the time. We cannot use the subsequent success of smallpox vaccinations as evidence that early proponents were always correct and protests unwarranted, since the protests may have been an important political signal that makes it clear that negative side effects will not be tolerated. Without them, given the history of medicine, negative side effects may have been more tolerated, as an unfortunate negative consequence of an ultimate good, as a result of junk utilitarian ideology. Fear of flight is a clearer example of the importance of rude protest and emotional impulses, since it is fear of flight that generates the market for safe flight. Without fear of flight there is no reason to believe air transport would be as safe as it is today, since industry regulation could not compete with lower cost alternatives in the black market (which doesn't exist precisely because of low risk tolerance) if the market didn't create the space for industry regulation.
I don't have any of Scott (J's) books to hand here, or any of the surrounding literature, so I'll just mention a couple of things from my own experience.
I don't think we're talking here about knowledge vs judgement. Technocrats, at least in Europe, are people who are credentialed, but may actually be totally ignorant about the subjects they claim to advise on. You'll have seen this if you've ever worked in an organisation that has been management consulted: in general, the people working in the organisation have a far better grasp of its strengths and weaknesses than outsiders can, but I've seen outside "technocrats" ruin and destroy organisations on the basis of complete ignorance about how, say, government works.
If you want to see technocratic arrogance at its worst, try the multibillion international "aid" and "development" industry, on which Scott (J) commented, but which has been comprehensively dismantled by the Cambridge economist Ha-Jon Chang, and more recently by Susan Woodward. Here, credentialed but ignorant technocrats from the IMF and World Bank get off planes and instruct governments what to do. The classic was in Yugoslavia in the 80s when the IMF demanded a radical decentralisation of the economy, not realising that the Yugoslav economy was already one of the most decentralised in the world. This led directly to the political tensions that caused the war. You see it in softer areas as well: a typical sequence of events (I've seen it) runs as follows:
-Hello, I am paid by a human rights programme funded by the EU. I have a Master's degree in Human Rights Law from the University of Goteborg and I have just finished an internship at the Swedish Foreign Ministry. In my bag I have a copy of various Swedish laws which you will adopt, and I will reform your justice system.
- But we already have highly complex codes of justice which may not have been written down but which we have been using for hundreds of years. You don't even speak our language.
-Tough.
It's not fair either to characterise this as a struggle between expertise and judgement. For example, traditional agriculture is evidence-based. It has to be, or you die. It's modern "technocratic" farming methods that are judgemental and normative. Same goes for traditional medicines: they represent an n=the largest number you can imagine experiment over thousands of years. Doctors, in my experience, are far less practical and more ideological in their approach (though I'm sure Scott A isn't).
Are the decision makers insulated from the consequences of their decisions, and from the facts on which their decisions should be based?
If only rich people are involved in making decisions (ie usually), they are insulated.
If only men are involved in making decisions, they are insulated.
If the people making decisions look only at statistics, they are ignoring the worms-eye view (and vice versa).
If the people making decisions look only at mathematical models, they are ignoring not only the "unquantifiable issues", but all the incorrectly quantified parameters.
If people who work with their hands are excluded from decision making, things won't go well for them.
It's a bit funny that examples (1)-(5) aren't so clear unqualified successes of evidence-based reasoning. Note also that none of them come with any form of evidence of success like a cost-benefit analysis.
(1) I'll focus on the "mandatory" part - is it clear that this is necessary? After all, per the link, the "mandatory" part was effectively removed in England in 1898, and people tend to mistrust things when they are forced on them.
(2) While I am in no way for segregation, desegregation was not without reasonable critics. Hannah Arendt was one, her main argument if I remember correctly was "top down decisions shouldn't be borne by the children".
(3) I have no opinion on this.
(4) Is it clear that the "scientists" know what they are doing, or would survive a cost-benefit analysis?
(5) Given that Scott has indicated in the comments that he doesn't want this to be a focus, I'll limit myself to saying that (a) in my experiences, lockdowns, initially widely popular, were more of a reaction to the people asking for them, the account there reads like a revisionist history and (b) this seems like a clear example of failure to me, at least by the cost-benefit calculations I've seen.
Also: I don't know what "compact polygons" are, did you mean to write "convex polygons".
Minor correction, which I haven't checked but am almost certain is correct. The nine Supreme Court judges who ordered the revamp of the school systems would have had JDs from Harvard and Yale, not PhDs.
As someone who works in the vicinity of government policy being made (I realise that's incredibly vague but anonymity requires it) I think part of the problem is that good technocracy is expensive and with lots of government policy we don't get it. We instead get rushed work by bureaucrats without the time and resources to properly assess the problem, or what is pushed by a particular special interest.
E.g. let's say you want to put together the new legislation regulating the production, sale and distribution of widgets. In an ideal world you'd get input from everyone affected by the problem, business, consumer groups, a representative sample of the population of widget users, etc. And come up with competing models for how you regulate them, (widget licencing, vs widget usage taxes, etc) and have experts compare the merits of the different model and trial them in smaller scale settings. Unfortunately that would require a lot of expensive staffing and takes several years. When there is political demands to have an immediate solution to the problem.
So what actually happens? If nobody really cares much about widgets, it gets delegated down the chain to a mid ranked person in the department of widgetology who puts together a plan based on the very limited time and information they have access to. And that maybe gets reviewed by other tired and overworked people before being passed back up the chain for political approval.
Or if its a politically important policy (your politicians have made lots of statements about the importance of the widget industry, and/or the harms of widget related pollution) you get a list of what the widget lobby wants, and what politicians think will be popular, and someone (possibly the same bureaucrat but often a twenty something political appointee) tries to make a coherent policy solution out of that set of probably incompatible demands. That gets bounced back and forth with politics and public opinion and maybe a few tweaks are made, then it is either rolled out or doesn't happen at all depending on political winds.
Unfortunately good policy development is slow and boring and happens in the background. So no politician ever loses out for cutting the funding for faceless bureaucrats in back rooms. And the benefits of better policy are long term and diffuse, so don't have a lot of political capital behind them.
I think that the fatal flaw of this essay is that it presents a false dichotomy between objectivity/rationality and subjectivity/values. Yet this doesn't actually exist.
Experts almost never limit themselves merely to what they can prove objectively. Their values are nearly always reflected in their advice, models, goals, etc. A good example of this is the military. The military is typically quite good at making military decisions and politicians that micromanage the military don't have a very good record. Yet no nation, other than military dictatorships, gives the military free reign. The reason is that if you put the military in charge, the decisions don't merely reflect military expertise, but also military values, which tend to deviate a lot from what the majority wants.
The goal of the system should be maximize the benefit from expertise, while minimizing the subjective values that tag along with it. After all, experts don't deserve to get their desires met more than anyone else, in a society where we hold everyone's desires to be of equal value (the idea behind democracy).
This is only in part a balancing act. In large part it is about proper checks & balances and incentives. We should reward experts when when they act on their expertise and correct/punish them if they act on their values (at least, if they don't align with the democratic or individualist values, depending on the context). This view on democracy is represented by: 'Throw the bastards out.'
It's not the job of the voters to select find and empower the rulers. It is the job of rulers to prove themselves worthy. It is the job of journalists earn the trust of the populace. The same for scientists. For companies. For banks and other financial institutions. Etc.
The attacks on technocracy, the elites and such aren't because people oppose reason or science, but because they feel that the game is rigged and their own desires aren't being given the fair weight in decision making by those with exceptional power.
PS. The more people's values diverge in society, especially between those with exceptional power and the rest of society, the harder it is to make the former do what the latter considers to be fair.
This is a brilliant post, and it neatly organizes some reactions I have had to the James C Scott most ardent supporters. So I will finally write a response to the review.
I would like to reverse the roles and add bit of personal experience *in favor* of "technocracy". I happened to have lived in 3 of the cities mentioned in the review of Seeing Like a State: Chicago, Rio and Brasilia. I think that it gets wrong views on all these cities.
First, Chicago. How can we assess the relative success of Chicago and Bruges? I would argue that it would be based on their capacity to improve lives. By that metric, Chicago is several times more successful than Bruges. Founded only in 1833, it attracted millions of previously impoverished people which were able to improve their lives significantly. The city was able to relatively effortlessly grow by orders of magnitude in a century. Bruges, by contrast, seems to only have some 100k people after a thousand years of existence. I don't doubt that standards of living in Bruges are better than Chicago, but are people improving their lives by moving into Bruges? This should make us look at Bruges as possibly the ultimate NIMBY city.
Now Rio and Brasilia. I am from Rio and moved to Brasilia at a point in my life. If I'm biased, I'm biased in favor of Rio where I had a happy childhood and early adulthood, and was able to enjoy everything that people most love about the city. But Rio is an extremely difficult city. Traffic, violence and a mountain on every corner makes it a very hard city to navigate.
Brasilia, however, is known for having some of the best quality of life in Brazil. I definitely can see why an American visiting Rio and Brasilia would absolutely prefer Rio, but living is different from visiting. And the fact is that if you talk to people who grew up in Brasilia, they absolutely love the city and trash talk Rio all the time. My hypothesis is that Brazilians are much more attached to their home city than Americans. Back in the 60s and 70s, when the first impressions about Brasilia were being formed, everybody in Brasilia was not from Brasilia. So they all missed their hometown and trash talked the new capital. Now that we have some generations of brasilienses, they all seem to love the city. It is a very expensive city which is maintained by federal tax dollars, much like DC. But from a Brazilian standpoint, Brasilia is great.
Note to Scott: I can see the Hidden Open Thread in the Top articles list at the bottom of this page below the comments, despite not being subscribed. I mention this because I know you were trying to make them invisible. Add another thing to Substack's to-do list.
Does Weyl still consider his most important work to be Quadratic Voting, a formal mechanism for making democracy better by measuring strength of preference? If so, the essay critiqued here would seem to be a cynical ploy to pitch his work to a new audience in new terms. Last I heard, a few years ago, he was struggling to convince economists and political scientists of its value.
There is another alternative to letting one political party draw up districts for its own benefit, which is giving control to " wise people who truly understand the state and its complex needs draw districts that group naturally-related areas together and make sure everyone has an equal say."
And, contra what Scott (Alexander) wrote, "But somehow whenever we ask our wise-people-who-truly-understand-the-state to do this, they always come up with weird pipe-cleaner shapes that vote exactly 51% Republican." - that isn't what happens. We Draw the Lines (.ca.gov) operates exactly on that basis and California has pretty good districts - +4 Democratic on PlanScore's metric on the US House, 0% on the state Senate, +2 Democratic on the state House. This compares, to say, North Carolina being +24 Republican, Ohio +15 and the one genuinely gerrymandered Democratic state (Maryland) at +11.
Not only that, there are other countries that have single-member districts and need boundaries (many countries don't have districting because they use proportional systems). There's Britain, there's Canada, there's Australia, there's France. They exclude partisans from the decision-making process and end up with districts without a strong bias to either party.
I am grateful for your taking the time to respond. There is a lot there to respond to and in general I think the exchange speaks for itself. However, I think there are few points where clarification is important for the exchange to be productive, which I'll briefly address here.
1. Let me start with concessions. There are many points where @slatestarcodex correctly highlights various areas where my grasp of beliefs and facts are limited or wrong, especially in the depth of my grasp of the views of the rationalist community. I freely admit that there are serious limits to how much I've been able to research the views of people in this community and I certainly hope they are not as I characterized them, though as I will point out below many elements of Scott's response confirm my concerns.
2. Given the last point, I fully acknowledge the danger of throwing stones lest I shatter my own glass house. However, leaving aside any blame, I think to make sense of my piece and Scott's response requires a bit of context that clearly he lacked. First, outside some specific blog post I wrote, I am best known as a mechanism designer. To cast me as a general opponent of technology and mechanisms runs against literally everything I am known for and have worked on my whole career. The piece was as much a self-critique and caution about taking the sorts of work I do in the wrong, over-zealous spirit as I had seen many in the rationalist community doing as anything else. Second, if responding to anything directly, it was this review of my book by https://www.aeaweb.org/articles?id=10.1257/jel.20191533 I think that piece perfectly exemplifies the spirit I am responding to and critiquing. I think we need to avoid that spirit of mechanism design.
4. Furthermore, the positive examples of technocracy @slatestarcodex refers to are...surprising. Two examples. To call school desegregation a technocratic invention papers over decades of community activism for desegregation. Perhaps even more dramatically looking at the coronavirus as an example of the success of technocracy runs against pretty much any reasonable reading of the international data. Danielle Allen and I have a piece coming
5. The last part of the piece is explicitly about the role I see mechanism playing in a democracy. I find it hard to understand how one could see the piece as opposed to mechanisms. My argument was that the appropriate way for mechanisms to be adopted
Is through public communication across lines of difference and in different value systems/communicative modes. One thing I find striking in the history of technology is that the vast majority of technologies that are actually useful today were pioneered by people who had similar critiques to mine here of technocracy, while those who zealous defend technocratic approaches have generally either not themselves actually developed successful technologies or have great technological dreams that have generally led to poor social outcomes. Consider Douglas Engelbart, Norbert Wiener, Jaron Lanier, etc. Calling people like this, most of whom were not even willing to express their views in the rationalistic terms I wrote in, "anti-technology" redefines technology to be only rigid and inhuman systems that fail. The process of socio-technological change has a far greater element of the "socio" when it succeeds than those focused on autonomous "technology" allow. Communication and collaboration outside of affordances of the technology itself are always critical to success. See, for example, Don Norman's Design of Everyday Things, or anything else in the field of human-centered design.
6. I think @slatestarcodex's insistence on breaking apart mechanisms v. judgement from top-down v. bottom-up misses a key part of the argument and of what sociologists of science have long said. There is no unitary thing called "science" or "mechanism". There are a variety
of disciplines of information processing across academic fields, across cultures, across communities with in a culture, etc. "Mechanism" is just how one group of people seeks to claim that their mode of reasoning is uniquely unbiased and unaccountable to other ways of
processing information. It is precisely this move, the unwillingness to think, speak or justify oneself on terms acceptable to those who think differently from you, that his response manifests and that concerns me.
7. A particularly striking example of this was his identification of "democracy" with the one-person-one-vote rule, an identification I have in the past been guilty of (hence the self-critique). This is not coterminal with what democracy means to most people, nor how most
political scientists think of it. I will not belabor this here, but I think it is a nice illustration of how much one's views can be narrowed by only looking through a "mechanistic" lens.
You know, it's funny, but I feel like this post exactly captures my struggle when it comes to infant sleep. Maybe I'm so sleep deprived that all I can think about is sleep these days, but I still do see it! Especially the first axis, the top-down v. bottom-up element. There are a lot of books out there on how many hours babies should sleep at night, how many naps they should be taking each day, how much awake time they should be given between each nap, etc. etc. These books are written by MDs and PhDs, the experts. On the other hand, there are my elderly female relatives yelling in my ear about how "books about babies are ridiculous! they sleep when they're tired! when I had kids, no books, just instinct, the way humans always did it, etc. etc."
Sometimes I feel like a reluctant technocrat. I impose the rules the experts tell me to impose but then my kid wakes up at 2:30am anyway and I'm tempted to throw them all out. I would love to have instead an algorithm for baby sleep!
Hey Scott, so I actually recently read Seeing Like A State, and had an argument with my doctor friend about decision making. We landed on the idea that the amount of like, written accepted knowledge about a thing factors heavily into whether intuition or rationality will be better, because rationality is only as good as maybe what you can read on books and blogs, but intuition works with knowledge people might have but can't spell out yet. So in realms like being a doctor and making doctor decisions that have been studied a ton, rational algorithmic scientific decision making might be better, but as a forester from 1910, maybe you try to analyze the data you've collected and you're worse than the farmer guy who isn't Science based because Science isn't at all advanced in your field yet. Both science/rationality and intuition are GIGO, but maybe it's easier to fool oneself with numbers.
Another facet we talked about with all this, and this is definitely a little bit more vague, is that using intuition lets the decision maker somehow use illegible data. Maybe this data is much less worthwhile than the hard data when it comes to diagnoses in some situations, but maybe it is actually worth something in others? He painted a scenario where he had to make a doctor decision where the available indicators made it a tossup decision, and I suggested that maybe in a tossup decision scenario, talking to the patient and like, listening to their breathing or something, might allow some sort of benefit. My point with this was that like yeah rational scientific decision making is good, but maybe lots of professions take it too far, even doctors.
The nugget of Weyl's essay that feels really timely to me right now was where he discusses how it is better to have a simple, understandable policy than it is to have a complex "optimal" policy. I think we can see that right now with covid vaccinations, where attempts to carefully prioritize who gets it become so convoluted that normal people can't understand what's going on. Then the punishment for failure scares hospitals enough to slow everything way down.
The question though: are the technocrats the smart folks on twitter screaming that we need to simplify vaccine administration, or are the technocrats the democratic governors coming up with unnecessarily complicated formulae? I've always tended to associate the term technocrats more with the first group, but maybe that's wrong - it's not like Nate Silver actually has any political power other than yelling at people on twitter. But usually when I hear technocrats used as a term of abuse, it feels like it's aimed at the same people who right now are trying to simplify vaccine administration, not the people who are trying to make it more complicated.
But in terms of actual policy, it does seem like making sure normies can understand the policy is a good practice for policymakers to follow.
If I just read the opening of Weyl's piece here, my first inclination is this is obviously wrong. Why would we not want experts using formal methods informed by data to be making important policy decisions?
I believe the actual answer to this is itself data-driven. Democracy, libertarianism, and populism, in somewhat true forms that probably don't exist, all spring from the observation that giving power to technocrats works only by luck in the short run and is guaranteed to fail in the long run because eventually whatever offices of power they hold are going to be captured by people driven by ideology or personal ego, not data. This isn't a critique of technocracy. It's a critique of allowing singular institutions to accumulate too much power at all. I don't really want a legislature that can be held hostage by whoever supports Marjorie Taylor Greene, but it's better than the alternative of a legislature that can't be held hostage by any means at all.
I don't know what the solution to this problem is. We want and I think need buy in and participation from as broad a swath of everybody as possible when making decisions that impact everybody. This means you can't exclude people that are driven by really stupid reasons. Is there a way to make a world with fewer of these people that doesn't devolve into Gulags, genocide, and eugenics? At its heart, this seems to what rationalism is trying to do, very slowly spread brain worms to individuals until enough of them are more rational and less ideological. At what point, if ever, is this expected to yield measurable social benefit? Trillions of years from now isn't a very satisfying answer to most people.
If you look at the specific failure cases you're citing here, they don't seem technocratic to me at all. Robert Moses was largely successful. Where his projects ended up having bad side effects, he seems to have been driven by racism, not data. Soviet and Chinese famines caused by centrally planned agriculture was driven by Marxist-Leninist ideology, not by data.
China right now actually seems to have a near all-time high sweet spot for a totalitarian regime in terms of delivering mostly good stuff for most of its population, achieving both better productivity and efficiency than more democratic governments, but it remains subject to the long run failure mode of what happens when the CCP is run by someone less like Xi Jinping and more like Kim Jong-un?
Just a quibble, but Scott's discussion of school desegregation is pretty lacking. In fact, it was not effective to have the Supreme Court, even a unanimous one, demand school desegregation, and the federal government did not really put its muscle behind the movement until after passage of the Civil Rights Act of 1964. Which is to say: going through a process that was better capable of demonstrating mass popular support and giving Southerners a full chance to argue their cause, only to lose decisively, was in fact much better at changing the world for the better than just hoping that the declaration of 9 judges could do so.
I'm not sure American constitutional law really fits into the "technocracy" debate, anyhow.
Reviewing all of the examples cited in this post against the ones from Seeing Like A State, maybe the distinguishing characteristic is that policies are non-technocratic if they become popular afterwards and technocratic if they don't (a position I am suggesting only partially in jest).
Or, to use Strong Bad's Rule of Awesome, "Too much of a good thing is an awesome thing, but too much of an awesome thing is... really dumb, and bad."
There's a tendency to erase the role of technocracy/statism/modernism/whatever you want to call it in things that have become successful, while pointing it out whenever it fails. Pruitt-Igoe failed, and obviously those nasty technocrats were imposing their vision on an urban populace that didn't want it. Levittown succeeded, and obviously this is a bottom-up free market success... never mind that Levittown was planned from the top down by Levitt & Sons and imposed upon a formerly rural area with the endorsement of governmental authorities. It was copied nationwide as other local governments changed their zoning to allow Levittown-style suburbs while making it hard to build anything else. And certainly a car-centric suburb of NYC could never work without the expansive highway networks built by one Robert Moses.
There may be a distinction to be drawn here but I'm not sure what it is.
Suggestion for Scott, when you mention examples of Gerrymandering, consider the pros and cons are of blaming Republicans instead of blaming the party in charge. The only pros I can think of is to signal that you are on the side of the democrats and you want to more accurately describe which party is more at fault. The cons are you annoy your Republican readers, and you minimize the fact that many Democrat states also use Gerrymandering, like Maryland and Illinois, just not quite to the same extent as Republicans.
Ending Gerrymandering could be a non partisan issue. Republican voters like to believe in fairness. That's why things like restoring felon voter rights passed by over 60% or the vote in Florida the same election that DeSantis one. Republican leaders are good at couching policies they like as being more fair. They claim voter ID laws will prevent illegal votes to be counted to gain support. I'm a Republican, and I hope one day we can form a non partisan coalition to end Gerrymandering in every state, even though Gerrymandering personally benefits me, since I live in a Republican state. But to build support for this, Gerrymandering should be couched as a non-partisan problem when parties in power unfairly minimize the votes of minority parties, not couched as Republicans oppressing Democrats.
Just something to consider Scott. Gerrymandering is just one specific example, but there are going to be a lot of times when you will have to make the judgement call of using both sides language or it's my outgroup's fault language. I think there are times when both are appropriate, I just think Gerrymandering is a good time to use both sides language.
Commenting before finishing reading this article so I might end up editing/deleting this shortly. Two things:
1) What do you mean 'Nobody ever defends technocracy'. A significant majority of the people I associate with very explicitly advocate for technocracy, even as (ironically enough) most of them have substantial disagreements on what that would actually be
2) _Most_ of your seeing-like-a-state proposed case studies are actually really bad and do not prove the point that sometimes central planning is good:
> School Desegregation
Despite peoples' well-meaning desire to fight racism, this was a major step along the way to centralized federal control of public education, and centralized federal control of public education is the single biggest reason why US schools are dysfunctional.
> The interstate highway system:
The interstate highway system itself is pretty great. But a side-effect of its development is that now almost all state roads are built with federal funding. This is a gigantic disaster, and you can read about this in much more detail than I could state here at http://strongtowns.org. tl;dr: federal government pays for road construction but not road maintenance. Road maintenance is a problem 20 years from now when all current politicians won't be elected anymore, and so they all punt on it. The result is that most places in the US have too many roads that they can't afford to maintain
> Climate change
You've been reading your own blog posts' comment threads long enough to know everything there is to say on this subject. It is not _at all_ clear that the governmental responses to climate change are unmitigated positives
> Coronavirus lockdowns
Anyone who seriously thinks that the specific coronavirus lockdowns we actually had in the US were good is both ignorant of all available data, and malicious towards the lives of millions. I trust that everyone here knows all of the relevant information and context so I won't add further
> By technocracy, I mean the view that most of governance and policy should be left to some type of “experts”, distinguished by meritocratically-evaluated training in formal methods used to “optimize” social outcomes
> Did you notice none of Weyl's examples of technocracy fit this definition at all? Robert Moses had zero formal training in urban planning or anything related to city-building. The Soviet leadership wasn't "meritocratically chosen".
Every attempt at technocracy will end the same way. It is all fine and good to say "if only the experts were in charge" but now you have a meta-problem of "how do you select experts?". It is not at all obvious who the experts are in any given field. Let a field self-select its experts and you enable nepotism and cliquishness. Create a standardized evaluative procedure and you will always be 5-10 years behind the cutting edge. And, no matter what mechanism you pick, people concerned with gaining and exercising power _will_ subvert your system and not only use it to their own ends, but use the fig leaf of 'we're the experts' to justify their decisions and quell dissent
I agree with many of the overall points, but I felt like the examples at the beginning entirely undermined everything that came afterward:
1. Mandatory vaccinations: Many European nations eschew mandatory vaccinations and have higher vaccination rates than the US. I'm not certain the mandatory part of this program is necessary, or even helpful. Especially as it feeds conspiracy-nut movements that make high vaccination rates (and implementation of new vaccines) more difficult. As an immunologist, I'm constantly having to answer objections arising from misunderstandings about vaccines. The people who object often cite mandatory programs and aggressive vaccination schedules as the impetus for their skepticism. Both of these features, incidentally, are technocratic inventions. Neither seems to be well-calibrated or necessary to achieve the desired result.
2. School desegregation: This example feels like a case of authoritarian vs. authoritarian. School desegregation was enacted against the state governments enforcing the policy - in some cases contra the will of the plurality of public opinion. It's a complicated story that doesn't fit neatly into a pro/anti-technocracy box. Maybe it's better to say it's technocrats vs. technocrats?
("How could you say the racists were technocrats?!" The theory at the time was that racial mixing would be bad for people of both races; that minorities were less capable than whites, and that mixing the two groups together would cause bad economic and social outcomes for white and minority communities. The theory was wrong, of course, but that's exactly the complaint against technocracy: it's based on prevailing ideas at the time it's enacted, not on an ideal of absolute Truth.)
3. Interstate highways: Is this really a win for the technocrats? It seems more like a much better case study than Brasilia of the bad unintended consequences that ensue when technocrats get involved. As I understand it, the interstate highway program helped to destroy the passenger rail system in the US (which had been more extensive than in Europe). It encouraged environmentally destructive urban sprawl, which in turn created many poor downstream effects, like increased congestion as everyone had to have their own cars, made mass/public transit difficult to implement, and destroyed millions of acres surrounding major population centers, which grew outward instead of upward. (Yes, the belt routes around major cities were part of the project. The thinking was that by encouraging urban sprawl this would make cities difficult to occupy by invaders - a great tactic for 19th century warfare built in the 20th century.) What would the US look like without the interstate highway project? It would probably have the great features of European cities technocrats have been complaining for years as being absent in the US - despite their best efforts to encourage them! (walkable city centers, viable mass transit, able to travel from place to place by modern high-speed rail)
5. Lockdowns: Were lockdowns really a strong move here? I'm sensitive to the argument that technocratic solutions are ideal for some problems, and it would seem like a pandemic such as this one would have been the best place to demonstrate that. Yet instead of good solutions based on solid epidemiological evidence, the technocratic solution was lockdowns - to the exclusion of many other, better solutions.
I wish we could have racked up a win here by implementing a different strategy. What if, after it became clear by early May who the vulnerable populations were, we implemented a cocooning strategy advocated by me and a bunch of others? What if we'd given free grocery and restaurant delivery to elderly people for a few months - at the cost of a few million dollars - to allow them to isolate more effectively? We could have paid people who'd already been infected (and recovered) to provide other services and helped high-risk people not be so lonely. It would have been a great example of how the technocrats got it right, implemented the obvious solutions based on good scientific evidence, saved lives, and kept the economy from tanking.
That's not what happened, though. Because the problem with technocracy is that it has to contend with Public Choice Theory, and it rarely comes out the victor in that exchange.
As an economist, I've generally appreciated Weyl's criticisms of technocracy, but I ultimately don't find them extremely useful from an EA perspective. EA's are now mostly interested in ensuring the long-term future (10,000 + years) goes well and because future generations don't exist yet, it is hard to see how 'local knowledge' considerations could come into play.
In fact, I see this as part of the case for long-termism. Because future people can't vote or participate in the market, their preferences are for the most part ignored. If there was a way to incorporate these preferences into democracy and markets today, I would be very happy! But, I don't really see how that would be possible. Technocracy seems to be the only option.
> In light of the ignorance of typical political leaders and members of the general public, we might be tempted by the idea of rule by experts, as in Plato’s Republic. Unfortunately, when it comes to descriptive social theory, even the experts’ knowledge is unimpressive, as demonstrated recently by the social psychologist Phillip Tetlock. Tetlock conducted a fifteen-year study in which he collected tens of thousands of predictions from hundreds of political experts concerning matters within their areas of expertise (for example, would the economy slide into recession, would the Soviet Union survive, who would win the next Presidential election, and so on). Tetlock’s finding, in brief, was that the best experts did only slightly better than chance at predicting outcomes. When asked to assign probabilities to their predictions, experts proved systematically overconfident; for example, events predicted with 100% confidence happened less than 80% of the time.
Pretty surprising that neither Weyl's article nor your critique mention futarchy. I would expect this to be one of the first things brought up, given that:
-- you are both pretty strongly Robin-Hanson-adjacent
-- two very common and salient critiques of technocrats are that they either (a) have the wrong incentives to make good empirical judgments or (b) conflate empirical judgments and value judgments, fail to realize their training confers no advantage in making value judgments, and arrogantly shut out the common people from any voice in value judgments
-- so a system explicitly designed to incentivize technocrats to make good empirical judgments, while leaving value judgments in the hands of the common people, seems like a pretty obvious thing to consider when discussing how to deal with common failings of technocracy!
> All of this is too bad, because if Weyl was right, that would be a huge point in favor of rational methods. Imagine pointing to a community that did something wrong for years, then learned that thing was wrong, then admitted it and changed - and imagine thinking of it as a strike against that community's methods!
Doesn't this violate the conservation of expected evidence? Rationalism can't win both ways, so if this being true were a point for Rationalism, then it being false should be a point against it, Bayes-wise (the size of points may be different inverse to expectation, but it would still be against). Since I don't really see how it being false is a point against Rationalism, then it feels like that couldn't actually be a point for it. Is my Bayescraft off?
The things Weyl says about Rationalism and EA come across as confused.
e.g. "The effective altruism movement...seeks to maximize the efficacy with which charitable donations are directed using standard rationalist methods. It is a tight-knit community that strongly privileges rationalist approaches over all other forms of knowledge-making (such as from the humanities, continental philosophy, or humanistic social sciences) and tends to dismiss input not formulated in rationalist terms. The community also has a strong and explicitly stated view that its activities uniquely contribute to the achievement of 'the good': of their top five recommendations of most productive careers by a leading community organization, two suggest being a researcher or support staff within the movement, and two others recommend working on the AI alignment problem (see the next point)." This sounds mildly like "Beware, EAs are an elitist cult who doesn't listen to outsiders". Like much of the article, it's subtle jabbing couched in academic language in order to mask its intellectual vacuousness.
"Until recently, much of the analysis and funding emerging from the community has pointed towards a focus on extremely unlikely but potentially catastrophic risks, such as alien, asteroid or biological catastrophes." Granted it was only 3 years ago that I pledged to give 10+% of my income to charity for the rest of my life, but I missed the memo about Our Focus On Aliens And Asteroids.
"Yet, interestingly, the conclusions of the analysis emerging from the community increasingly undermine these foci and the approach of the community more broadly. In particular, recent research in the community suggests that the greatest and most probable risks to be avoided are anthropogenic" Um, hello? AI catastrophic risks *are* anthropogenic! So are the "biological catastrophes" we worry about! What, did you think we thought that a naturally-occuring virus was likely to wipe out humanity in the next 100 years?
"Leaders in the community have in turn suggested that the most effective ways to avoid these are likely finding solutions to problems of political organization and legitimacy of social systems to help reduce the likelihood of conflict or inability to cooperate in the provision of critical global public goods." Actually, yes. Very astute. Unfortunately, Weyl doesn't recognize that the ability of the community to change its mind is arguably its greatest strength. On Tim Urban's "Thinking Latter", this is the main difference beteween endless conflict between echo chambers, and productive cooperation and understanding.
"Ironically, those “political” goals, social reforms, and public confidence building are precisely the sorts of activities that effective altruists have long viewed as 'non-rigorous' and ineffectual." Oh really? Please quote me any well-known community member ever claiming that these things were "ineffectual"?
"Worse, the extremely elitist, segregated, and condescending approach to philanthropy encouraged by the community has created widespread public backlash..." Um, what backlash? I think if there's a backlash it'll be generated by people like yourself who are happy to portray it inaccurately, in much the same way as there was a "backlash" against having a "socialist" president who "wasn't born in America" and "backlash" against a "fascist" president held in check only by "The Resistance".
This is a well-written essay and it brings up very interesting arguments. I also wish I hadn't read it. I try to stay away from culture war topics as much as possible, because when they do end up finding me, they occupy my attention much more than they deserve.
A filter option would make me very happy, to keep receiving posts about the history of amphetamines, but not essays about culture war topics I otherwise would not have found. (A bit like https://xkcd.com/2071/). Yes, I'm aware I'm wishing for a stronger filter bubble, even though filter bubbles have been found to be the source of all modern evil.
Why is "I̶t̶'̶s̶ ̶n̶o̶t̶ ̶a̶ ̶s̶t̶o̶r̶y̶ ̶t̶h̶e̶ ̶J̶e̶d̶i̶ ̶w̶o̶u̶l̶d̶ ̶t̶e̶l̶l̶ ̶y̶o̶u̶" in a notably different (sans serif, for starters) font than the rest of the piece?
You mind find the recent fiasco with awarding UK A-level and GCSE grades last year an interesting case study:
:- It was not possible to hold exams this year, but we still had to award qualifications in some way, and this couldn't be postponed because of e.g. university places.
:- The government decision was that teachers should make, and announce, predictions of the grades their students would receive, and then an algorithm would be used comparing those grades against grades of previous years at the same institution (and possibly previous grades of the same students?) to normalise those.
:- In ordinary years teachers release predicted grades too, and those grades are consistently higher on average than the actual grades turn out to be.
:- When the algorithm awarded many students lower grades than the unprocessed predictions from their teachers, there was outcry. The government was forced to back down and a week or two later the original, unnormalised grades were awarded.
:-The outcry was exacerbated by the fact that for very small classes it was felt that there wasn't enough data to normalise, and so the raw, probably-too-high grades were used in those places only; this turned out to disproportionately benefit private schools with children of rich parents at them.
:- This has put universities in a terrible bind. Universities typically make initial offers to students early in the year - "we will definitely take you if you get two A's and a B" - and then, if not enough students make their demanded grade (which will be the case for a lot of universities each year) they take additional students once the grades are announced, through a process called clearing. This year, clearing happened /before/ the government backed down, all places were filled, and then it turned out that a bunch more students had also made their predicted grades, and places had to be found for them to, which will cause significant overcrowding problems for the next three or four years.
:- It's probably all going to happen again this year.
Having read through the original article, I do think Weyl has a point on (some) discussion of AI alignment, a lot of discussions really are, I quote: "a bit like posing the “genius dictator alignment problem”: how can we ensure that if there is a brilliant dictator, he will serve the interests of the broader public?"
I think the solution is going to involve limiting the power that computers have over people and ensuring they remain accountable (probably through the use of other AIs). Humans are the ones developing AI, this really is just the ancient question of how to limit the power some people have over other people.
Long term, I think I worry more about the human alignment problem than the AI alignment problem - which human values will we align AI to?
The idea is that policy should "do what is says on the tin": make it easy for everyone, regardless of whether they agree with you, to understand what it is say you're going to do, and then to see whether or not you actually succeeded in doing that thing.
A simple policy that is less "optimal" or less fine-tuned may have better outcomes because its legibility makes it more likely to get the requisite political support. A meta-rational technocrat should be optimizing for policy legibility!
I am someone who is guilty of what you are criticizing, definitely. I appreciate this essay tremendously and definitely agree. Too many people want to use "technocracy" or some equal and opposite thing (let's say "irrationality") as their One Move That Ends the Debate, when at the end of the day you have to get into the nitty gritty. As a humanist, I say, ah, there's no mechanism for avoiding that nitty gritty! As a chastened one, I add, I can't really find much fault in the nitty gritty you provide here, and also, I don't really think you or most notables in your community believe that mechanism must supplant human judgment everywhere and anywhere. At any rate, very glad to see you writing again so I can read essays like this one.
"In the Languedoc there is a vineyard that teaches us an important lesson about textbook learning and its application to the world. In the early Seventies it was bought by a wealthy couple, who consulted professors Emile Peynaud and Henri Enjalbert, the world’s leading academic oenologist and oenological geologist respectively. Between them these men convinced the couple that their new vineyard had a theoretically ideal microclimate for wine-making. When planted with theoretically ideal vines whose fruits would be processed in the optimal way according to the up-to-date science of oenology, this vineyard had the potential to produce wine to match the great first growths of Bordeaux. The received wisdom that great wine was the product of an inscrutable (and untransferable) tradition was quite mistaken, the professors said: it could be done with hard work and a fanatical attention to detail. The couple, who had no experience of wine-making but much faith in professorial expertise, took a deep breath and went ahead.
If life were reliably like novels, their experiment would have been a disaster. In fact Aimé and Véronique Guibert have met with a success so unsullied that it would make a stupefying novel [...] The first vintage they declared (in 1978) was described by Gault Millau as ‘Château Lafite du Languedoc’; others have been praised to the heights by the likes of Hugh Johnson and Robert Parker. The wine is now on the list at the Tour d’Argent and the 1986 vintage retails at the vineyard for £65 a bottle. The sole shadow on the lives of these millionaires is cast by the odd hailstorm."
Good post. Simultaneously encouraged and disappointed that the most controversial example in the comments is not the culture war-adjacent example, but instead whether strong mitigation was needed for COVID-19.
Please, stop using Brasília as an example of a failed technocratic project. Niemeyer didn't build a Modernist village, because he built no village at all - he merely designed some (very popular) public buildings and spaces. Brasilia (and a good deal of the Federal District) was projected by Lucio Costa. I lived there for 3 months, and most of the residents (specially family people) love it. And there was no bottom-up alternative, since there was nothing in its place. What really bothered me about the city is some peculiar governance issues (though I can't say other Brazilian capitals, like São Paulo antes Rio, fate better) and the artificial environment provided by it's status as the federal capital city.
The three examples you give; mandatory vaccination, desegregation and the federal highway system; prove the point you are trying to disprove. The heavy handed all or nothing approach that the federal government took in desegregating schools resulted in a devastating backlash, a backlash that is at the root of how the Republican party got coopted by extreme Libertarians (see Democracy in Chains). What should have been done is they needed to start with desegregating the teachers first.
When Japan and Europe were investing in trains, the US built freeways. What's happened? We have a world that revolves around the personal automobile with horrendous environmental consequences and the endless nightmare of congested roads and marathon commutes and we are stuck with it. Talk to anyone in their eighties. They all remember the freeways being shoved down their throats. The freeways ruined American cities. Gutted them.
As for mandatory vaccination, I don't actually know much about that only that the most basic human instinct is freedom. As soon as you try and control behavior or actions, you are asking for trouble. The really simple solution always was and will be: Wrong thing hard, right thing easy. Vaccination is free and efficient, if you don't want to be vaccinated, you have to pay a yearly fee. $500? That should do it. That would ensure so few people remain unvaccinated that it wouldn't be a concern. Better yet, give people $10 for getting vaccinated. You'd be surprised how little money you have to pay people to ensure their co-operation.
A debate that seems to coming up often is whether farm collectivisation counts as a genuine technocratic act if it wasn't actually scientifically valid. The problem is that the authoritarian governments that implemented it would of course have claimed that their policies were scientific; except that all the scientists were ideologically compromised by the state and members of the public couldn't see the data. So perhaps one component of a true technocracy is open science - the studies motivating a policy decision should be, as far as possible, visible to and experimentally verifiable by the public.
I haven't read this thread, so apologies if this is redundant, but many of the distinctions you mention, along with many additional examples, were discussed in The Future and its Enemies back in 1998. (BTW, Technocracy was actually a political movement in the 20th century, with the idea that all of society and economy should be centrally directed along the lines with which total war had been waged.)
In my wording, technocracy is an ideology fundamentally opposed to spontaneous orders that emerge from the interaction of "private" or "autonomous" choices by subsystem actors; those subsystem actors can be as rationalist and expert-driven as they choose, and the technocrat will still oppose their execution and (unplanned) interaction. Moreover it insists that virtue inheres in the mere intention to predict and control the output of the total system, and whether the effort is successful according to some objective function is of secondary importance. There are technocrats who want to subject system planning to majoritarian control, those who want it delegated to a coterie of experts of some sort, etc.
But it is a category error IMO to apply the term "technocracy" to something operating at a subsystem level--say, a college admissions office, or the quality-control department of a factory--embedded in a larger competitive and cooperative set of entities whose actions it must take as given. Designing an entire city from scratch, or subsuming an entire agricultural sector into collective farming, sure. The "-cracy" aspect should be taken seriously as referring to political control over a whole society, not to the private endeavors of actors within the system.
"The Soviet leadership wasn't "meritocratically chosen". "
What is 'meritocracy' Scott, and what evidence do you have that they weren't chosen in a meritocratic manner? Similarly, is the supposed 'failure' of Soviet collective farming something you actually studied using objective agricultural data, or is it just a vague feeling you have?
Funny. I was just reading about how Taiwan keeps discussions on an online policy forum productive, by allowing likes but instead banning replies.
I thought that was an interesting contrast to the comment moderation discussion last week, where SSC prefers to ban likes and encourage replies.
The Taiwan system also includes some other features I would have called "technocratic," to nudge people towards civil, productive discussion.
So I come on to mention it, only to find it already under discussion as an example contra technocracy. Highly unexpected!
Given my confusion, I agree there might be a gap in definitions here.
I still think it's worth reading about that system on its own merits, completely separate from the way it's being referenced in this discussion. One can appreciate developments in Taiwan without taking a firm stance on whatever either side is ultimately calling "technocracy."
Without commenting either way on the broader argument here, just wanted to point out that some of your examples in the beginning are odd. How is school desegregation an example of technocratic rule? It has been the result of intense grassroots pressure from the civil rights movement, preceded by a civil war. Stranger still to describe it as an attempt to get schools to more closely adhere to the Constitution, which was written by slave owners who couldn't have imagined such a thing. Also fail to see how scientists sounding the alarm about global warming are "technocrats". The scientific community hasn't called the shots in the (failed) struggle against climate change, and our technocratic rulers have miserably failed us.
Wow this post was incredible. A couple of years back somebody recommended slate star Codex to me and I read a few posts and determined that Scott was just a horrible writer with no ability to structure or organize his thoughts in a clear concise way. but here he has done that so masterfully that I have to assume I just somehow misjudged him before.
I was actually a co-founder of a nonprofit that advocates for a political reform called approval voting, and it got its legs after many years of struggle due to a grant from the open philanthropy foundation led by the effect of altruism movements William McCaskill. one of the other founders of that non-profit is a Princeton math PhD who created a redistricting algorithm called the shortest split line algorithm. I wonder if that's one of the ones Scott is thinking about in this post.
On a bit of a tangent, that nonprofit recently brought approval voting to St Louis, the second US city to use it after starting in Fargo.
Because of this post I went back and reread the Seeing Like a State review. The review was written in 2017 (around when I read the book myself); in 2019 I got big into Roman history for personal reasons (dorkiness).
And I gotta say, it feels like the big thing missing from this conversation is Imperial Rome. The Romans loved big round numbers and big tiled rectangles. Their fundamental statecraft innovation was taking a bunch of people from deeply embedded rich local cultures, dragging them thousands of miles away, making them march around in rectangles, build perfectly rectangular forts, and then eventually mustering them out to perfectly rectangular farm allotments in unrelated parts of the empire—farms so rectangular you can still see the rectangles etched across the face of Europe today.
And it worked LIKE CRAZY. Through the magic of rectangles, Rome took itself from a relatively minor city state marginal to both the Greek and Punic worlds to the predominant social order of the Mediterranean basin. And, as far as I can tell—though I'm far from a scholar—Rome mostly didn't experience the failure modes associated with, e.g., Soviet collective farms or Prussian scientific forests. I don't hear, for example, that the veteran farms were plagued with collapse or underproduction.
Now, of course, RECTANGLES(TM) are not the whole story. As with many successful empires Rome relied on all sorts of rich syncretism of assimilated and affiliated cultures. Auxiliary forces fought in the armies; foreign gods were smushed willy-nilly into the pantheon; and so on. So it's certainly not a pure triumph of High Modernism two millennia early.
But maybe this puts a bit of a different spin on the Love of Rectangles and its sundry failures. The fact of the matter is that Rectangle Magic, properly applied, is extremely powerful—not just in theory but in practice. And not every use of Rectangle Magic yields Rectangle Rot. And to get the gains out of Rectangle Magic you kind of have to bet big on it and commit in a big way—you get the legions out of a bunch of technocratic reforms and top-down dictates.
So maybe the High Modernists were kinda . . . on to something? Or rather, they were investing in a strategy that had paid out before: take some domain of human life which is currently a bit of a mess and ADD RECTANGLES UNTIL IT CRIES OUT FOR MERCY. Which has a track record other than failure!
Makes me kind of wish that there was a companion book to Seeing Like a State about the record of success and growth associated with rationalization, from the state's perspective. You could call it Seeing Like a State, maybe.
"This isn't just true for colleges - we know that giving everyone IQ tests and letting the top scorers into gifted programs ends up with better representation of gifted minorities than letting teachers use their judgment."
This isn't correct, see the paper:
"We test this hypothesis using data from a unique natural experiment conducted by a large and diverse school district in the state of Florida (hereafter “the District”). State law dictates that students must achieve a minimum of 130 points on a standard IQ test to qualify for gifted status. English language learners (ELLs) and free-or-reduced price lunch (FRL) participants are subject to a lower 116 point threshold, known as “Plan B” eligibility."
"Climatologists created complicated formal models to determine how quickly global temperatures might rise"
I apologies because this is about climate modelling, and nitpicking your off-the-cuff example, but I thought I saw an opportunity to say something useful so here goes:
So there is a big and important misconception here: climate models are NOT currently (properly considered) part of the scientific case for global warming; and people saying they are, convinces a lot of climate-change-skeptics that they are on a firmer footing than they actually are -- a lot of climate-change-skepticism is pointing out limitations in model reliability. And their prominence in the news may fuel climate-skepticism (when people generally learn to treat similar concepts like economic-models and forecasts with friendly contempt).
But this is irrelevant because the scientific case around global warning stands independent of modelling.
The go-to source for evaluations of the effectiveness of state-of-the-art climate-models is the IPCC (the UN body tasked for compiling evidence on climate change) regular reports on "The Physical Science Basis", each of which contains a section "Evaluation of Climate Models" (or some paraphrase). They're all available online and very readable. Go check out.
The 4th edition of their report (a few years back) devotes 74 pages to this topic. The newest report is a short refresher on this pointing out which things have progressed since then.
I guess one way of looking at it is that the "aggregate climate model" the IPCC uses, is the aggregate of predictions of many different models, many of which have completely different assumptions about dominant mechanisms. They use this aggregation because some models are good at some things and some good at other things. But I think it's difficult to build a scientific case on an aggregate of contradictory scientific cases?... not like "well, if you believe X, then Y, and if you believe not-X, then also Y, so Y" -- that's a fine argument -- but this argument would be "I think X and I think not-X, and their aggregate is Y". They are a best effort at prediction. They aren't a case. The rest of the 1007 page report (AR4) is the case.
I have a bone to pick with all of your examples. I think they are all interesting and subtle and not so clearcut. The interstate highway system is clearly a good thing and marvel of engineering, but many beautiful neighborhoods were paved, and it ended up creating the problems Not Just Bikes is always complaining about. As for school desegregation, I don't know much about this, but didn't busing and integration also come with seriouses increases in crime? Mandatory vaccination seems like a good thing for smallpox, but there is a serious debate about COVID vaccine mandates, given vaccine side effects. As for climate change, I am almost certain that any climate legislation will do much more harm than good. The climate movement is completely deranged, anti-nuclear, calling for a radical transformation that if seriously pursued would be disastrous. And the COVID lockdowns were according to Bhattacharya the greatest public health disaster in history.
History is written by elites, so whatever they do (and did) is wise and good and certainly not a big mistake.
I can't tell if the inclusion of Covid Lockdowns with everything else is parody or not (especially given Scott's prior scratchpad post on QALYs...)
hrm, isn’t there a huge populist backlash against Gates? chips in the
vaccines and all that?
I wonder, to some extent, how much things like this are reflective of the ascendancy of an intellectual movement that had some real and rational critiques of what came before it but has more or less failed to establish a positive program of its own.
My wife has a masters in urban planning. The acquisition of said masters involved a great deal of, basically, self-flagellation on behalf of the discipline of urban planning, in which the students seemed to learn a lot more about the (genuine) ways in which urban planning royally fucked up the mid-20th century than about what it could do right using what we'd learned from that experience. (I exaggerate somewhat, but, like, not all that much.)
None of this was that shocking to me, because I'd gotten into Jim Scott in college in the 2000s and pushed him on friends like a cool new drug. High modernism wasn't exactly riding high in 2007 (although I guess the US government was actively attempting to summon a functional democratic capitalist Iraq out of military force and ultra-crystallized McKinsey so maybe I'm wrong) but nonetheless the critique was fresh enough that it grabbed our attention.
And so, you'd think, maybe out of this will come some new synthesized perspective, but it doesn't really seem like that happened. I'm a data guy now and I constantly emphasize the importance of being aware of what the metrics actually measure and steering clear of optimizing for misguided targets, so I've clearly been influenced, but the case you're presenting above sure sounds like a description of a critical movement that spiralled of joyously criticizing its defeated bogeymen rather than engaging with people who have internalized a few lessons from either direction.
I'm not averse to the argument but citing climate change as an example? Climate change is the poster child for technocratic over-confidence, wildly wrong predictions and the failure to balance costs with benefits.
Scott, I love your reasoning and think you have a great sense of which facts are important to a point, but when the facts *don't* matter, you almost invariably get them wrong. SCOTUS have JDs (or in the past LLBs) and Brown principally concerned the 14th Amendment, which was less than 90 years old at the time.
1) I'm confused about 1-5 - of these 1 and 4 seem objectively correct with later (scientific) knowledge, 2 seems to be correct in the sense that it's a no-brainer under our current values, 3 and 5 seem like a good examples of the state messing things up (lockdowns aside, the state *also* said "no masks" there for a long time).
2) A much bigger example of something that bothered me in Seeing Like a State - he tears up a lot of western agricultural practices that failed to generalize, but wasn't the green revolution a thing that happened, largely successfully, mostly eradicating third world hunger? What was different about it from the failure cases he's quoting? And most importantly, why does Scott never ask this? It reminds me of the three number pattern game in HPMOR - Scott assumes he has a theory to explain the failure of some big projects, but he never tests his theory on any government successes to see if it can separate them.
I was going to dismiss Weyl out of hand, as I do for anyone who uses phrases like "strongly privileges rationalist approaches". But I realize that I can fully consider what he's saying and still dismiss him, based on the axis I think really matters: Chosen Vs. Coerced.
But I want to first take a whack at a similar issue that I think Scott elides in his introduction: "Nobody ever defends technocracy. It's like "elitism" or "statism". There is no Statist Party. Nobody holds rallies demanding more statism. There is no Citizens for Statism Facebook page with thousands of likes and followers."
Here I think we need to reach for our Orwell, and his Politics and the English Language: "https://www.orwell.ru/library/essays/politics/english/e_polit"
Consider for instance some comfortable English professor defending Russian totalitarianism. He cannot say outright, ‘I believe in killing off your opponents when you can get good results by doing so’. Probably, therefore, he will say something like this:
‘While freely conceding that the Soviet regime exhibits certain features which the humanitarian may be inclined to deplore, we must, I think, agree that a certain curtailment of the right to political opposition is an unavoidable concomitant of transitional periods, and that the rigors which the Russian people have been called upon to undergo have been amply justified in the sphere of concrete achievement.’"
For the record, I can give arguments against all 5 examples Scott mentions, and, yes, I can drag in quotes from black activists who said that forced integration was a fiasco, but if I explain what's wrong with all of the different initiatives we'll be here all day.
The point is that euphemism is always a way to evade what one is talking about.
Anyway, back to Weyl - the contradiction here is that he is against the E.A. movement which is purely voluntary. No one is forcing anyone to take part in it, and - as I understand it - everyone can make the case and present the evidence for and against it, and for and against what causes the E.A. movement tackles.
If you read the article, you see he somehow conflates both Stalinist central planning and economic liberalization. So the whole thing is a con job - it's saying: "Some people who thought they knew better advocated for _this_, and other people who think they know better advocated for _that_, so they're the same, and you shouldn't listen to either - because I know better!"
I don't think that technocracy vs "unnamed good way of making decisions" is necessarily a good or useful dichotomy, but I do think that there is value in looking at what various solutions or ways of doing things are optimizing for. For instance, school admissions. Though this seems a little like a methodology debate - which I think is what talks about technocracy really are, discussions on the best way to do things - it seems more deeply like a debate about what you want in a student body. If you are optimizing for a student body who can expose each other to a diversity of backgrounds and life experiences, your admissions process is going to look different than if you are just attempting to get the most intelligent students.
I think a real postmodernist or critical theorist or whatever you want to call them would say that power is what is missing from both your and from Weyls analysis though. Just because something isn't data driven or centrally planned doesn't mean it's less likely to perpetuate power differentials or intergenerational wealth or elite capture, it's just that "technocracy" gives things a veneer of objectivity that would otherwise be missing, which gives it more staying power than otherwise. At least, that's what I think of as the actual critique of technocracy - not that it necessarily has worse outcomes than other methods but that it pretends to be above the entrenched interests etc.
It’s very easy to agree with your essay in general, but counting the handling of COVID as a triumph for technocracy [outside of China] undermines the fundamental premise. The mask fiasco, the absurd discussions surrounding border closures, the idiotic conservatism surrounding vaccine approval, much of the school debate, etc were mostly elite failures. A good dose of Folksy Common Sense wouldnt have been terrible in many of those debates.
I don't know what this is a product of, but I do not have negative connotations for the word 'technocrat'. The main usage I associate it with is when, say, The Economist says Country X has appointed a technocrat as Minister of Salad Forks or whatever, i.e. someone chosen for their competence rather than affiliation to a particular party; it's invariably presented as a good thing.
Does the word have universal negative associations in other media circles, or just in particular ones that Weyl would be a part of?
"it worries me that everyone analyzes the exact same three examples of the failures of top-down planning: Soviet collective farms, Brasilia, and Robert Moses."
Excuse me, but Robert Moses is an example of a SUCCESS of top-down planning!
N.B. Since it appears to be impossible to find replies after I've been informed of them by email (or even to find my original comment), I'm going to put off responding to such until the site is fixed. In the meantime, I'll just sit here and pontificate.
Technocrats were more useful when 1) the world had much less information and 2) that information was only available to technocrats pre-Internet.
Today it's impossible for technocrats to absorb even a sliver of the world's information and they're often not much smarter than a hobbyist or blogger somewhere.
Technocrats also tend to impose highly centralized decision-making and solutions, when decentralized systems are more robust and experimental.
Since you mentioned Brazil, here's an example of the technocrats in Brazil saving the country from hyperinflation:
https://www.npr.org/sections/money/2010/10/04/130329523/how-fake-money-saved-brazil
The really strange thing about the Interstate is that it was very likely heavily inspired by a nightmarish journey across the continent that Dwight D Eisenhower made when he was a young Lieutenant Colonel - one that must have made him think "There has to be a better way to do this!" - and it's hard to get more common-man and less elitist than "Army man made big remembers his American road trip"!
(https://www.eisenhowerlibrary.gov/research/online-documents/1919-transcontinental-motor-convoy)
(https://www.fhwa.dot.gov/publications/publicroads/03mar/05.cfm)
(Granted, during World War II Eisenhower was a general who had never seen combat, so he may be a disgraceful example of an elite holding positions of unwarranted high rank; but he did fairly well for himself despite his lack of experience.)
The only thing I can really think of here that maybe can be added is the idea of technocracy as the bureaucratization of government. Notably, I don't think their existence is a harmful features of technocracy, but rather that the process vs judgement is associated with principle-agent harms.
Which one I think is harmful depends on whether we're considering internal decision making or external relations. Internally, we see that having democratic systems of judgement (ie. commitees decide FDA approval vs simple rules or housing by right vs housing permit systems) adds complexity and delay, as well as undermining good outcomes.
I think the flip side is that having process based external relations vs democratic based external relations undermines responsiveness, especially if combined with internal judgement. It means that the personnel is unresponsive to outside feedback loops, which makes it hard for voters to ensure their best interests are being taken care of (eg vs. military contractors in the military or working towards liberal goals under R admin). The combination that I see as 'bad technocracy' is internal judgement (as opposed to formal process), which makes policy illegible for lawmakers, alongside external process (vs democracy) that makes it hard to effect personnel.
I don't think it's an accurate characterization to describe Weyl as taking the "judgment" side of the "mechanism vs judgment" axis! After all, he is a strong advocate for quadratic funding, Harberger taxes (aka COST aka SALSA...), superlinear quantity subsidies and other "radical mechanism design ideas". That said, he's certainly not anti-judgement in any sense either; I view him as being in the "we need the mechanisms, and we also need stronger public discourse to help us better use the mechanisms" camp.
I think the best way to view his position on this issue is that he has a strong opinion on the tradeoffs between different *types* of mechanisms. This piece co-written by him and myself is IMO a good explainer of the core idea:
https://vitalik.ca/general/2018/11/25/central_planning.html
Basically the idea is that mechanism designers should explicitly have simplicity and explainability as core values to a much greater degree than (math/econ) academics are naturally inclined to accept today. A more concrete example of what he's getting at is this piece from him:
https://promarket.org/2020/05/28/how-market-design-economists-engineered-economists-helped-design-a-mass-privatization-of-public-resources/
That piece talks about his experiences in spectrum auction design, and how a very complicated auctioning style actually was abused and led to a few actors getting spectrum licenses at very unfairly cheap prices. Furthermore, the "illegibility" of the mechanism makes it inherently hard to prove to the public that something wrong happened, because defenders of the outcome can just spout some technobabble that goes way over the heads of not just the average person but even the smartest person that the average person listens to.
Both markets + property rights, and democratic voting, are far superior on this dimension. The mechanisms are simple enough that even lay people can clearly understand how the mechanisms are supposed to work and so they trust them. (Yes, I know, lots of people express suspicion about both property rights and voting all the time. But I'm sure you can agree that if people's cars and houses and who the president is were instead allocated using "theoretically optimal" VCG auctions, then things would be 1000x worse).
Furthermore, markets and voting are both not *opposed* to "judgement"; it's pretty clear that in both cases, they perform reasonably well precisely because a lot of complicated and (necessarily) illegible judgements are being made by the participants in the mechanism! Though clearly this is not true of all mechanisms; Twitter likes/retweets, for example, have no way to express concepts like "I think this tweet is socially harmful and should be shown to fewer people, though it's not quite so egregious that it should be removed outright".
I do disagree with parts of his EA critique. Particularly, it's worth noting that one of the projects more prominently favored by the EA community, GiveDirectly, involves *just giving poor people money and trusting them to do what they want with it* - in many ways the ultimate anti-technocratic form of charity. (Though EA does do its fair share of more-centrally-planned things, like handing out anti-malaria bednets, as well).
I think there might be a mistake in: "Some people analyzed a bunch of data, came up with an algorithm for diagnosing heart attacks, and told doctors they should use the algorithm instead of their own judgment. (...) The interesting part (search "Goldberg Rule" in that link) is what happens when you give doctors the algorithm and tell them to use it to supplement their judgment." - I just looked up "Goldberg rule" in the article linked and it seems to be about diagnosing psychosis, not heart attacks.
First I cannot tell you how overjoyed I am that you are writing regularly again. The mental clarity is sensational.
One nit-pick - the first time we hear about Scott is
"These axes prove their use when we use them to analyze Weyl's vs. Scott's conceptions of legibility."
I only learnt from the comments you were talking about someone called Jim Scott, and not referring to yourself in the third person.
There are multiple Statist parties, two of which spend billions of dollars evangelizing each election season.
Also, during my undergraduate years, I once came across a bound copy of several issues of the Technocracy Party newsletter when I was supposed to be studying. They were a Vietnam-era group which seemed to be pushing for evidence-based socialism to be applied by scientific-minded leaders - probably would have been big Yang fans.
Great article! One thing on
'Suppose you're the FDA trying to regulate cigarettes, and your doctors and bureaucrats have come up with a plan. What do you do now to "seek democratic feedback"? Put a suggestion box outside the FDA office?
Here in the UK - yes. There's a requirement to have a public consultation on new policy - anyone can answer. It's not a vote but you have to consider the arguments raised and if you just ignore things raised without addressing them you can get into legal trouble.
Strangely, Scott misses the opportunity to discuss the large scientific literature that actually tests whether "mechanism" predictions (and thus decisions) are better than "judgments" predictions, or what researchers call algorithms, or actuarial, or model predictions vs. human, clinical predictions. Fortunately, I wrote a blogpost summarizing this evidence a few years ago (2016), so readers can just read that.
https://emilkirkegaard.dk/en/2016/07/clinical-vs-statistical-prediction/
TL;DR formal methods essentially always beat human judgments, even when given the same data, even when humans are given the predictions from the formal model. I interpret this in line with human's desire to be in control, which is apparently stronger than their desire for good outcomes. In these cases, it seems wiser to set strong defaults in favor of whatever the models are saying, and let some minority of humans opt out of that if it's not too much of a problem (maybe for a price).
I think more discussions on policy should draw attention to the fact that policy disagreements often hinge on different preferences (disagreements on facts seem to spring from a desire to support one's preferences). In this light, there is obviously no such thing as objectively correct policy, unless everyone already agrees on the desired outcome.
The issue with technocrats is that they often seem to think everyone agrees with their own policy goals, so all that's left is convincing everyone that their proposal for achieving this goal is optimal.
It’s interesting that one of the primary benefits of systematic technical decisions is legibility, when many systems being implemented by the state are increasingly illegible (either proprietary, or a result of inscrutable ML models that even the creators can’t explain). I wonder if Weyl would agree that technical decision makers actually *are* more legible. (Not disagreeing with the general point that scrutable >> inscrutable, just pointing out how that classification is getting less-broadly-true over time)
Others have pointed to this, but I want to present specific arguments as to why 4 out of the 5 case studies at the beginning aren't particularly strong cases.
1) Vaccines - I will skip this because I agree it's a good example of productive "technocracy", and it's the strongest of the 5 case studies.
2) Desegregation - The high court mandating an end to segregation was good. But the technocratic court-ordered implementation left much to be desired. By the 1970s, busing had an under 10% approval rate among all races. Just about everyone thought - and I believe most would still agree today - that as technocratic federal courts further embroiled themselves in the day-to-day planning of desegregation, without respect for local considerations, the worse matters became.
3) Interstates - The case study cites how mob/activist input caused highway plans to be changed from technocratically-determined direct routes to ones that caused less turmoil. Is the argument that these changes were bad? In my eyes, this case study shows mob/activist input causing a potentially-brutal technocratic policy to be changed to something more humane. So I don't think this is technocratic success; it's at best evidence of the success of a mixed approach. (Also, I think your discussion of it thus far doesn't sufficiently engage with the real pain that can come with the government deciding to take and destroy your home.)
4) Climate Change - I'll give you cap-and-trade, but that's not the most significant technocratic anti-AGCC policy at all. Most technocratic proposals against global warming have been a wash at best, or actively harmful. Natural gas is better than the status quo, and its rise is one of the biggest reasons US emissions are decreasing, but climate-conscious technocrats want to ban fracking because "fossil fuels bad". Technocratic consensus is for the Paris accords, despite simple back-of-the-envelope calculations demonstrating that they impose *nowhere near* strong enough burdens on e.g. China and India to have any significant impact. And here in California, climate-conscious technocrats chose gasoline taxes as the ballot issue to champion, despite most economists' recognition that gas is an inelastic good and this will not significantly lower emissions. Maybe certain smart people on Twitter have the solutions all figured out. But the technocrats aren't listening to them, and what they're actually doing in the present day is not helping.
5) Lockdowns - I don't want to dwell on this because I'm sure others will. I'll just say - literally today, it was made clear that the state with the highest death rate (and death count) in the country got there by intentionally ordering COVID patients to be placed in nursing homes, among the most vulnerable possible populations, and that they then covered it up. I know you probably wrote this before the news broke, but still, I think it might just give the lie to the idea that the technocratic response to COVID was a triumph.
"One Hedgehog Tool"? Is this a reference to something?
Small correction: Colin Powell has an approval rating one point higher than Gates (https://www.rasmussenreports.com/public_content/politics/favorables/major_political_figures).
A really good, nuanced critique of Weyl's essay, which quite annoyed me for exactly the reasons you describe above. Weyl in general seems to have gone backward over the last few years - I really enjoyed his book Radical Markets, but he recently withdrew support for some of the policies he recommended on grounds that to me seemed shaky and ideologically motivated.
Reading the comments, I'm noticing a lot of people conflating technocracy with centralisation; while this has been the case historically I don't see it as a general rule. Decentralised systems can absolutely be technocratic if they arise on the basis of rational and scientific principles. Bitcoin is a great example - the poster child of decentralised finance, based on a mathematical algorithm. In general, I'm very bullish about the application of the findings of complex systems science to governance. I think using principles listened from systems biology/sociology will allow technocrats to design systems that benefit everyone while avoiding the pitfalls of the old school high modernism. It needs a name - Systems Modernism, perhaps?
M-nci-s M-ldb-g and you both mentioned Russell conjugations on Substack in the same week.
Is that a coinci-- [remembers I'm talking to the author of Unsong] ...
> "humanities, Continental philosophy, and the humanistic social sciences" - isn't that usually code for stuff like queer theory, postcolonial theory, and postmodernism?
"code for" seems like a cheap and unnecessary jab... I would agree, however, that the humanistic social sciences, etc., etc., *consist mostly of* stuff like postmodernism, etc., etc.
I'm not a New Yorker, but it seems like Robert Moses got a lot of big things done that make NYC more livable for its immense population. I'm from the San Fernando Valley, which used to do big infrastructure projects like turning the LA River into a giant concrete ditch after the 1938 flood and building the Hollywood Freeway in 1955. But very little has been done since, despite a big increase in the number of households.
My guess is that Moses' early infrastructure projects had high returns, but after awhile they had diminishing marginal returns while Moses became more set in his ways as he aged.
Thus, the huge backlash against him.
But I suspect Moses's reputation will rebound after Robert Caro is gone.
Hey, just so you know, Colin Powell got a 77% approval rating, just narrowly edging out Gates :)
I recommend you use James Scott’s full name. I kept thinking “Scott” meant you.
"Last time anyone checked, Gates had an approval rating of 76%, the highest of any figure asked about and literally higher than God."
This is slightly misleading; while Gates does have a higher *approval* rating than God (who had 52%), Gates' disapproval rating was 20% whereas Gods' was only 9%.
The entire premise of this post is wrong. Technocracy is not a Russell Conjugation, although it can be used as such. Technocracy is not evidence-based decision making, it is the creation of a veneer of complexity to avoid criticism and scrutiny of decisions, where only people who agree with the ultimate decision are allowed criticise the justifications for said decision. Its the facade of public scrutiny where none exists, because the people allowed to criticise will not upset the apple cart in the knowledge that they will lose their voice (and eventually their job security) by doing so.
Technocrats are not pilots taking us to new destinations is complex feets of engineering, they are scientistic frauds.
The incredibly basic differential equations and thumb in the air assumptions upon which they are based, calibrated early and unmodulated by subsequent reality, that were used to justify indefinite lockdowns are a clear example of this.
But even the smallpox vaccination protests may have been legitimate at the time. We cannot use the subsequent success of smallpox vaccinations as evidence that early proponents were always correct and protests unwarranted, since the protests may have been an important political signal that makes it clear that negative side effects will not be tolerated. Without them, given the history of medicine, negative side effects may have been more tolerated, as an unfortunate negative consequence of an ultimate good, as a result of junk utilitarian ideology. Fear of flight is a clearer example of the importance of rude protest and emotional impulses, since it is fear of flight that generates the market for safe flight. Without fear of flight there is no reason to believe air transport would be as safe as it is today, since industry regulation could not compete with lower cost alternatives in the black market (which doesn't exist precisely because of low risk tolerance) if the market didn't create the space for industry regulation.
If the people in charge suck, it doesn't matter which rulebook they're using.
I don't have any of Scott (J's) books to hand here, or any of the surrounding literature, so I'll just mention a couple of things from my own experience.
I don't think we're talking here about knowledge vs judgement. Technocrats, at least in Europe, are people who are credentialed, but may actually be totally ignorant about the subjects they claim to advise on. You'll have seen this if you've ever worked in an organisation that has been management consulted: in general, the people working in the organisation have a far better grasp of its strengths and weaknesses than outsiders can, but I've seen outside "technocrats" ruin and destroy organisations on the basis of complete ignorance about how, say, government works.
If you want to see technocratic arrogance at its worst, try the multibillion international "aid" and "development" industry, on which Scott (J) commented, but which has been comprehensively dismantled by the Cambridge economist Ha-Jon Chang, and more recently by Susan Woodward. Here, credentialed but ignorant technocrats from the IMF and World Bank get off planes and instruct governments what to do. The classic was in Yugoslavia in the 80s when the IMF demanded a radical decentralisation of the economy, not realising that the Yugoslav economy was already one of the most decentralised in the world. This led directly to the political tensions that caused the war. You see it in softer areas as well: a typical sequence of events (I've seen it) runs as follows:
-Hello, I am paid by a human rights programme funded by the EU. I have a Master's degree in Human Rights Law from the University of Goteborg and I have just finished an internship at the Swedish Foreign Ministry. In my bag I have a copy of various Swedish laws which you will adopt, and I will reform your justice system.
- But we already have highly complex codes of justice which may not have been written down but which we have been using for hundreds of years. You don't even speak our language.
-Tough.
It's not fair either to characterise this as a struggle between expertise and judgement. For example, traditional agriculture is evidence-based. It has to be, or you die. It's modern "technocratic" farming methods that are judgemental and normative. Same goes for traditional medicines: they represent an n=the largest number you can imagine experiment over thousands of years. Doctors, in my experience, are far less practical and more ideological in their approach (though I'm sure Scott A isn't).
I would add another axis:
Are the decision makers insulated from the consequences of their decisions, and from the facts on which their decisions should be based?
If only rich people are involved in making decisions (ie usually), they are insulated.
If only men are involved in making decisions, they are insulated.
If the people making decisions look only at statistics, they are ignoring the worms-eye view (and vice versa).
If the people making decisions look only at mathematical models, they are ignoring not only the "unquantifiable issues", but all the incorrectly quantified parameters.
If people who work with their hands are excluded from decision making, things won't go well for them.
It's a bit funny that examples (1)-(5) aren't so clear unqualified successes of evidence-based reasoning. Note also that none of them come with any form of evidence of success like a cost-benefit analysis.
(1) I'll focus on the "mandatory" part - is it clear that this is necessary? After all, per the link, the "mandatory" part was effectively removed in England in 1898, and people tend to mistrust things when they are forced on them.
(2) While I am in no way for segregation, desegregation was not without reasonable critics. Hannah Arendt was one, her main argument if I remember correctly was "top down decisions shouldn't be borne by the children".
(3) I have no opinion on this.
(4) Is it clear that the "scientists" know what they are doing, or would survive a cost-benefit analysis?
(5) Given that Scott has indicated in the comments that he doesn't want this to be a focus, I'll limit myself to saying that (a) in my experiences, lockdowns, initially widely popular, were more of a reaction to the people asking for them, the account there reads like a revisionist history and (b) this seems like a clear example of failure to me, at least by the cost-benefit calculations I've seen.
Also: I don't know what "compact polygons" are, did you mean to write "convex polygons".
Minor correction, which I haven't checked but am almost certain is correct. The nine Supreme Court judges who ordered the revamp of the school systems would have had JDs from Harvard and Yale, not PhDs.
As someone who works in the vicinity of government policy being made (I realise that's incredibly vague but anonymity requires it) I think part of the problem is that good technocracy is expensive and with lots of government policy we don't get it. We instead get rushed work by bureaucrats without the time and resources to properly assess the problem, or what is pushed by a particular special interest.
E.g. let's say you want to put together the new legislation regulating the production, sale and distribution of widgets. In an ideal world you'd get input from everyone affected by the problem, business, consumer groups, a representative sample of the population of widget users, etc. And come up with competing models for how you regulate them, (widget licencing, vs widget usage taxes, etc) and have experts compare the merits of the different model and trial them in smaller scale settings. Unfortunately that would require a lot of expensive staffing and takes several years. When there is political demands to have an immediate solution to the problem.
So what actually happens? If nobody really cares much about widgets, it gets delegated down the chain to a mid ranked person in the department of widgetology who puts together a plan based on the very limited time and information they have access to. And that maybe gets reviewed by other tired and overworked people before being passed back up the chain for political approval.
Or if its a politically important policy (your politicians have made lots of statements about the importance of the widget industry, and/or the harms of widget related pollution) you get a list of what the widget lobby wants, and what politicians think will be popular, and someone (possibly the same bureaucrat but often a twenty something political appointee) tries to make a coherent policy solution out of that set of probably incompatible demands. That gets bounced back and forth with politics and public opinion and maybe a few tweaks are made, then it is either rolled out or doesn't happen at all depending on political winds.
Unfortunately good policy development is slow and boring and happens in the background. So no politician ever loses out for cutting the funding for faceless bureaucrats in back rooms. And the benefits of better policy are long term and diffuse, so don't have a lot of political capital behind them.
I think that the fatal flaw of this essay is that it presents a false dichotomy between objectivity/rationality and subjectivity/values. Yet this doesn't actually exist.
Experts almost never limit themselves merely to what they can prove objectively. Their values are nearly always reflected in their advice, models, goals, etc. A good example of this is the military. The military is typically quite good at making military decisions and politicians that micromanage the military don't have a very good record. Yet no nation, other than military dictatorships, gives the military free reign. The reason is that if you put the military in charge, the decisions don't merely reflect military expertise, but also military values, which tend to deviate a lot from what the majority wants.
The goal of the system should be maximize the benefit from expertise, while minimizing the subjective values that tag along with it. After all, experts don't deserve to get their desires met more than anyone else, in a society where we hold everyone's desires to be of equal value (the idea behind democracy).
This is only in part a balancing act. In large part it is about proper checks & balances and incentives. We should reward experts when when they act on their expertise and correct/punish them if they act on their values (at least, if they don't align with the democratic or individualist values, depending on the context). This view on democracy is represented by: 'Throw the bastards out.'
It's not the job of the voters to select find and empower the rulers. It is the job of rulers to prove themselves worthy. It is the job of journalists earn the trust of the populace. The same for scientists. For companies. For banks and other financial institutions. Etc.
The attacks on technocracy, the elites and such aren't because people oppose reason or science, but because they feel that the game is rigged and their own desires aren't being given the fair weight in decision making by those with exceptional power.
PS. The more people's values diverge in society, especially between those with exceptional power and the rest of society, the harder it is to make the former do what the latter considers to be fair.
This is a brilliant post, and it neatly organizes some reactions I have had to the James C Scott most ardent supporters. So I will finally write a response to the review.
I would like to reverse the roles and add bit of personal experience *in favor* of "technocracy". I happened to have lived in 3 of the cities mentioned in the review of Seeing Like a State: Chicago, Rio and Brasilia. I think that it gets wrong views on all these cities.
First, Chicago. How can we assess the relative success of Chicago and Bruges? I would argue that it would be based on their capacity to improve lives. By that metric, Chicago is several times more successful than Bruges. Founded only in 1833, it attracted millions of previously impoverished people which were able to improve their lives significantly. The city was able to relatively effortlessly grow by orders of magnitude in a century. Bruges, by contrast, seems to only have some 100k people after a thousand years of existence. I don't doubt that standards of living in Bruges are better than Chicago, but are people improving their lives by moving into Bruges? This should make us look at Bruges as possibly the ultimate NIMBY city.
Now Rio and Brasilia. I am from Rio and moved to Brasilia at a point in my life. If I'm biased, I'm biased in favor of Rio where I had a happy childhood and early adulthood, and was able to enjoy everything that people most love about the city. But Rio is an extremely difficult city. Traffic, violence and a mountain on every corner makes it a very hard city to navigate.
Brasilia, however, is known for having some of the best quality of life in Brazil. I definitely can see why an American visiting Rio and Brasilia would absolutely prefer Rio, but living is different from visiting. And the fact is that if you talk to people who grew up in Brasilia, they absolutely love the city and trash talk Rio all the time. My hypothesis is that Brazilians are much more attached to their home city than Americans. Back in the 60s and 70s, when the first impressions about Brasilia were being formed, everybody in Brasilia was not from Brasilia. So they all missed their hometown and trash talked the new capital. Now that we have some generations of brasilienses, they all seem to love the city. It is a very expensive city which is maintained by federal tax dollars, much like DC. But from a Brazilian standpoint, Brasilia is great.
Note to Scott: I can see the Hidden Open Thread in the Top articles list at the bottom of this page below the comments, despite not being subscribed. I mention this because I know you were trying to make them invisible. Add another thing to Substack's to-do list.
Does Weyl still consider his most important work to be Quadratic Voting, a formal mechanism for making democracy better by measuring strength of preference? If so, the essay critiqued here would seem to be a cynical ploy to pitch his work to a new audience in new terms. Last I heard, a few years ago, he was struggling to convince economists and political scientists of its value.
"1. Mechanical district creation:"
There is another alternative to letting one political party draw up districts for its own benefit, which is giving control to " wise people who truly understand the state and its complex needs draw districts that group naturally-related areas together and make sure everyone has an equal say."
And, contra what Scott (Alexander) wrote, "But somehow whenever we ask our wise-people-who-truly-understand-the-state to do this, they always come up with weird pipe-cleaner shapes that vote exactly 51% Republican." - that isn't what happens. We Draw the Lines (.ca.gov) operates exactly on that basis and California has pretty good districts - +4 Democratic on PlanScore's metric on the US House, 0% on the state Senate, +2 Democratic on the state House. This compares, to say, North Carolina being +24 Republican, Ohio +15 and the one genuinely gerrymandered Democratic state (Maryland) at +11.
Not only that, there are other countries that have single-member districts and need boundaries (many countries don't have districting because they use proportional systems). There's Britain, there's Canada, there's Australia, there's France. They exclude partisans from the decision-making process and end up with districts without a strong bias to either party.
don't think the interstate highway system should be included on a list of things that made society unambiguously better
I am grateful for your taking the time to respond. There is a lot there to respond to and in general I think the exchange speaks for itself. However, I think there are few points where clarification is important for the exchange to be productive, which I'll briefly address here.
1. Let me start with concessions. There are many points where @slatestarcodex correctly highlights various areas where my grasp of beliefs and facts are limited or wrong, especially in the depth of my grasp of the views of the rationalist community. I freely admit that there are serious limits to how much I've been able to research the views of people in this community and I certainly hope they are not as I characterized them, though as I will point out below many elements of Scott's response confirm my concerns.
2. Given the last point, I fully acknowledge the danger of throwing stones lest I shatter my own glass house. However, leaving aside any blame, I think to make sense of my piece and Scott's response requires a bit of context that clearly he lacked. First, outside some specific blog post I wrote, I am best known as a mechanism designer. To cast me as a general opponent of technology and mechanisms runs against literally everything I am known for and have worked on my whole career. The piece was as much a self-critique and caution about taking the sorts of work I do in the wrong, over-zealous spirit as I had seen many in the rationalist community doing as anything else. Second, if responding to anything directly, it was this review of my book by https://www.aeaweb.org/articles?id=10.1257/jel.20191533 I think that piece perfectly exemplifies the spirit I am responding to and critiquing. I think we need to avoid that spirit of mechanism design.
3. Scott claims that critics of technocracy always critique precisely the same examples. This is odd, given that my essay has several examples outside that cannon. Did Scott not see these? I was blowing the whistle on one (promarket.org/2020/05/28/how…) at roughly the same time I wrote the technocracy piece. These are contemporary, not chestnuts, and conducted by precisely the circle (https://promarket.org/2020/05/28/how-market-design-economists-engineered-economists-helped-design-a-mass-privatization-of-public-resources/) whose condescending critiques of transparency and public engagement I was responding to.
4. Furthermore, the positive examples of technocracy @slatestarcodex refers to are...surprising. Two examples. To call school desegregation a technocratic invention papers over decades of community activism for desegregation. Perhaps even more dramatically looking at the coronavirus as an example of the success of technocracy runs against pretty much any reasonable reading of the international data. Danielle Allen and I have a piece coming
out on this (we were both deeply involved in developing a response plan here https://ethics.harvard.edu/Covid-Roadmap that significantly influenced now-President Biden's response), but perhaps the sharpest point here is that the country, Taiwan, which performed best in the virus was led in part by Audrey Tang who moved back to Taiwan after being immersed in and repulsed by the rationalist movement in Silicon Valley (see e.g. https://www.wired.com/story/how-taiwans-unlikely-digital-minister-hacked-the-pandemic/) and dedicated herself to doing things differently in Taiwan (see her amazing poetic job description here: https://twitter.com/audreyt/status/767953441746411524).
5. The last part of the piece is explicitly about the role I see mechanism playing in a democracy. I find it hard to understand how one could see the piece as opposed to mechanisms. My argument was that the appropriate way for mechanisms to be adopted
Is through public communication across lines of difference and in different value systems/communicative modes. One thing I find striking in the history of technology is that the vast majority of technologies that are actually useful today were pioneered by people who had similar critiques to mine here of technocracy, while those who zealous defend technocratic approaches have generally either not themselves actually developed successful technologies or have great technological dreams that have generally led to poor social outcomes. Consider Douglas Engelbart, Norbert Wiener, Jaron Lanier, etc. Calling people like this, most of whom were not even willing to express their views in the rationalistic terms I wrote in, "anti-technology" redefines technology to be only rigid and inhuman systems that fail. The process of socio-technological change has a far greater element of the "socio" when it succeeds than those focused on autonomous "technology" allow. Communication and collaboration outside of affordances of the technology itself are always critical to success. See, for example, Don Norman's Design of Everyday Things, or anything else in the field of human-centered design.
6. I think @slatestarcodex's insistence on breaking apart mechanisms v. judgement from top-down v. bottom-up misses a key part of the argument and of what sociologists of science have long said. There is no unitary thing called "science" or "mechanism". There are a variety
of disciplines of information processing across academic fields, across cultures, across communities with in a culture, etc. "Mechanism" is just how one group of people seeks to claim that their mode of reasoning is uniquely unbiased and unaccountable to other ways of
processing information. It is precisely this move, the unwillingness to think, speak or justify oneself on terms acceptable to those who think differently from you, that his response manifests and that concerns me.
7. A particularly striking example of this was his identification of "democracy" with the one-person-one-vote rule, an identification I have in the past been guilty of (hence the self-critique). This is not coterminal with what democracy means to most people, nor how most
political scientists think of it. I will not belabor this here, but I think it is a nice illustration of how much one's views can be narrowed by only looking through a "mechanistic" lens.
8. On the AI stuff, more here (https://www.wired.com/story/opinion-ai-is-an-ideology-not-a-technology/) and much more coming soon in a paper I am writing with Divya Siddarth
You know, it's funny, but I feel like this post exactly captures my struggle when it comes to infant sleep. Maybe I'm so sleep deprived that all I can think about is sleep these days, but I still do see it! Especially the first axis, the top-down v. bottom-up element. There are a lot of books out there on how many hours babies should sleep at night, how many naps they should be taking each day, how much awake time they should be given between each nap, etc. etc. These books are written by MDs and PhDs, the experts. On the other hand, there are my elderly female relatives yelling in my ear about how "books about babies are ridiculous! they sleep when they're tired! when I had kids, no books, just instinct, the way humans always did it, etc. etc."
Sometimes I feel like a reluctant technocrat. I impose the rules the experts tell me to impose but then my kid wakes up at 2:30am anyway and I'm tempted to throw them all out. I would love to have instead an algorithm for baby sleep!
Hey Scott, so I actually recently read Seeing Like A State, and had an argument with my doctor friend about decision making. We landed on the idea that the amount of like, written accepted knowledge about a thing factors heavily into whether intuition or rationality will be better, because rationality is only as good as maybe what you can read on books and blogs, but intuition works with knowledge people might have but can't spell out yet. So in realms like being a doctor and making doctor decisions that have been studied a ton, rational algorithmic scientific decision making might be better, but as a forester from 1910, maybe you try to analyze the data you've collected and you're worse than the farmer guy who isn't Science based because Science isn't at all advanced in your field yet. Both science/rationality and intuition are GIGO, but maybe it's easier to fool oneself with numbers.
Another facet we talked about with all this, and this is definitely a little bit more vague, is that using intuition lets the decision maker somehow use illegible data. Maybe this data is much less worthwhile than the hard data when it comes to diagnoses in some situations, but maybe it is actually worth something in others? He painted a scenario where he had to make a doctor decision where the available indicators made it a tossup decision, and I suggested that maybe in a tossup decision scenario, talking to the patient and like, listening to their breathing or something, might allow some sort of benefit. My point with this was that like yeah rational scientific decision making is good, but maybe lots of professions take it too far, even doctors.
The nugget of Weyl's essay that feels really timely to me right now was where he discusses how it is better to have a simple, understandable policy than it is to have a complex "optimal" policy. I think we can see that right now with covid vaccinations, where attempts to carefully prioritize who gets it become so convoluted that normal people can't understand what's going on. Then the punishment for failure scares hospitals enough to slow everything way down.
The question though: are the technocrats the smart folks on twitter screaming that we need to simplify vaccine administration, or are the technocrats the democratic governors coming up with unnecessarily complicated formulae? I've always tended to associate the term technocrats more with the first group, but maybe that's wrong - it's not like Nate Silver actually has any political power other than yelling at people on twitter. But usually when I hear technocrats used as a term of abuse, it feels like it's aimed at the same people who right now are trying to simplify vaccine administration, not the people who are trying to make it more complicated.
But in terms of actual policy, it does seem like making sure normies can understand the policy is a good practice for policymakers to follow.
If I just read the opening of Weyl's piece here, my first inclination is this is obviously wrong. Why would we not want experts using formal methods informed by data to be making important policy decisions?
I believe the actual answer to this is itself data-driven. Democracy, libertarianism, and populism, in somewhat true forms that probably don't exist, all spring from the observation that giving power to technocrats works only by luck in the short run and is guaranteed to fail in the long run because eventually whatever offices of power they hold are going to be captured by people driven by ideology or personal ego, not data. This isn't a critique of technocracy. It's a critique of allowing singular institutions to accumulate too much power at all. I don't really want a legislature that can be held hostage by whoever supports Marjorie Taylor Greene, but it's better than the alternative of a legislature that can't be held hostage by any means at all.
I don't know what the solution to this problem is. We want and I think need buy in and participation from as broad a swath of everybody as possible when making decisions that impact everybody. This means you can't exclude people that are driven by really stupid reasons. Is there a way to make a world with fewer of these people that doesn't devolve into Gulags, genocide, and eugenics? At its heart, this seems to what rationalism is trying to do, very slowly spread brain worms to individuals until enough of them are more rational and less ideological. At what point, if ever, is this expected to yield measurable social benefit? Trillions of years from now isn't a very satisfying answer to most people.
If you look at the specific failure cases you're citing here, they don't seem technocratic to me at all. Robert Moses was largely successful. Where his projects ended up having bad side effects, he seems to have been driven by racism, not data. Soviet and Chinese famines caused by centrally planned agriculture was driven by Marxist-Leninist ideology, not by data.
China right now actually seems to have a near all-time high sweet spot for a totalitarian regime in terms of delivering mostly good stuff for most of its population, achieving both better productivity and efficiency than more democratic governments, but it remains subject to the long run failure mode of what happens when the CCP is run by someone less like Xi Jinping and more like Kim Jong-un?
Just a quibble, but Scott's discussion of school desegregation is pretty lacking. In fact, it was not effective to have the Supreme Court, even a unanimous one, demand school desegregation, and the federal government did not really put its muscle behind the movement until after passage of the Civil Rights Act of 1964. Which is to say: going through a process that was better capable of demonstrating mass popular support and giving Southerners a full chance to argue their cause, only to lose decisively, was in fact much better at changing the world for the better than just hoping that the declaration of 9 judges could do so.
I'm not sure American constitutional law really fits into the "technocracy" debate, anyhow.
Reviewing all of the examples cited in this post against the ones from Seeing Like A State, maybe the distinguishing characteristic is that policies are non-technocratic if they become popular afterwards and technocratic if they don't (a position I am suggesting only partially in jest).
Or, to use Strong Bad's Rule of Awesome, "Too much of a good thing is an awesome thing, but too much of an awesome thing is... really dumb, and bad."
This is incredible, I am more impressed with the comments than I am with the article itself. This is what the internet should be like all over.
There's a tendency to erase the role of technocracy/statism/modernism/whatever you want to call it in things that have become successful, while pointing it out whenever it fails. Pruitt-Igoe failed, and obviously those nasty technocrats were imposing their vision on an urban populace that didn't want it. Levittown succeeded, and obviously this is a bottom-up free market success... never mind that Levittown was planned from the top down by Levitt & Sons and imposed upon a formerly rural area with the endorsement of governmental authorities. It was copied nationwide as other local governments changed their zoning to allow Levittown-style suburbs while making it hard to build anything else. And certainly a car-centric suburb of NYC could never work without the expansive highway networks built by one Robert Moses.
There may be a distinction to be drawn here but I'm not sure what it is.
Suggestion for Scott, when you mention examples of Gerrymandering, consider the pros and cons are of blaming Republicans instead of blaming the party in charge. The only pros I can think of is to signal that you are on the side of the democrats and you want to more accurately describe which party is more at fault. The cons are you annoy your Republican readers, and you minimize the fact that many Democrat states also use Gerrymandering, like Maryland and Illinois, just not quite to the same extent as Republicans.
Ending Gerrymandering could be a non partisan issue. Republican voters like to believe in fairness. That's why things like restoring felon voter rights passed by over 60% or the vote in Florida the same election that DeSantis one. Republican leaders are good at couching policies they like as being more fair. They claim voter ID laws will prevent illegal votes to be counted to gain support. I'm a Republican, and I hope one day we can form a non partisan coalition to end Gerrymandering in every state, even though Gerrymandering personally benefits me, since I live in a Republican state. But to build support for this, Gerrymandering should be couched as a non-partisan problem when parties in power unfairly minimize the votes of minority parties, not couched as Republicans oppressing Democrats.
Just something to consider Scott. Gerrymandering is just one specific example, but there are going to be a lot of times when you will have to make the judgement call of using both sides language or it's my outgroup's fault language. I think there are times when both are appropriate, I just think Gerrymandering is a good time to use both sides language.
Commenting before finishing reading this article so I might end up editing/deleting this shortly. Two things:
1) What do you mean 'Nobody ever defends technocracy'. A significant majority of the people I associate with very explicitly advocate for technocracy, even as (ironically enough) most of them have substantial disagreements on what that would actually be
2) _Most_ of your seeing-like-a-state proposed case studies are actually really bad and do not prove the point that sometimes central planning is good:
> School Desegregation
Despite peoples' well-meaning desire to fight racism, this was a major step along the way to centralized federal control of public education, and centralized federal control of public education is the single biggest reason why US schools are dysfunctional.
> The interstate highway system:
The interstate highway system itself is pretty great. But a side-effect of its development is that now almost all state roads are built with federal funding. This is a gigantic disaster, and you can read about this in much more detail than I could state here at http://strongtowns.org. tl;dr: federal government pays for road construction but not road maintenance. Road maintenance is a problem 20 years from now when all current politicians won't be elected anymore, and so they all punt on it. The result is that most places in the US have too many roads that they can't afford to maintain
> Climate change
You've been reading your own blog posts' comment threads long enough to know everything there is to say on this subject. It is not _at all_ clear that the governmental responses to climate change are unmitigated positives
> Coronavirus lockdowns
Anyone who seriously thinks that the specific coronavirus lockdowns we actually had in the US were good is both ignorant of all available data, and malicious towards the lives of millions. I trust that everyone here knows all of the relevant information and context so I won't add further
> By technocracy, I mean the view that most of governance and policy should be left to some type of “experts”, distinguished by meritocratically-evaluated training in formal methods used to “optimize” social outcomes
> Did you notice none of Weyl's examples of technocracy fit this definition at all? Robert Moses had zero formal training in urban planning or anything related to city-building. The Soviet leadership wasn't "meritocratically chosen".
Every attempt at technocracy will end the same way. It is all fine and good to say "if only the experts were in charge" but now you have a meta-problem of "how do you select experts?". It is not at all obvious who the experts are in any given field. Let a field self-select its experts and you enable nepotism and cliquishness. Create a standardized evaluative procedure and you will always be 5-10 years behind the cutting edge. And, no matter what mechanism you pick, people concerned with gaining and exercising power _will_ subvert your system and not only use it to their own ends, but use the fig leaf of 'we're the experts' to justify their decisions and quell dissent
I agree with many of the overall points, but I felt like the examples at the beginning entirely undermined everything that came afterward:
1. Mandatory vaccinations: Many European nations eschew mandatory vaccinations and have higher vaccination rates than the US. I'm not certain the mandatory part of this program is necessary, or even helpful. Especially as it feeds conspiracy-nut movements that make high vaccination rates (and implementation of new vaccines) more difficult. As an immunologist, I'm constantly having to answer objections arising from misunderstandings about vaccines. The people who object often cite mandatory programs and aggressive vaccination schedules as the impetus for their skepticism. Both of these features, incidentally, are technocratic inventions. Neither seems to be well-calibrated or necessary to achieve the desired result.
2. School desegregation: This example feels like a case of authoritarian vs. authoritarian. School desegregation was enacted against the state governments enforcing the policy - in some cases contra the will of the plurality of public opinion. It's a complicated story that doesn't fit neatly into a pro/anti-technocracy box. Maybe it's better to say it's technocrats vs. technocrats?
("How could you say the racists were technocrats?!" The theory at the time was that racial mixing would be bad for people of both races; that minorities were less capable than whites, and that mixing the two groups together would cause bad economic and social outcomes for white and minority communities. The theory was wrong, of course, but that's exactly the complaint against technocracy: it's based on prevailing ideas at the time it's enacted, not on an ideal of absolute Truth.)
3. Interstate highways: Is this really a win for the technocrats? It seems more like a much better case study than Brasilia of the bad unintended consequences that ensue when technocrats get involved. As I understand it, the interstate highway program helped to destroy the passenger rail system in the US (which had been more extensive than in Europe). It encouraged environmentally destructive urban sprawl, which in turn created many poor downstream effects, like increased congestion as everyone had to have their own cars, made mass/public transit difficult to implement, and destroyed millions of acres surrounding major population centers, which grew outward instead of upward. (Yes, the belt routes around major cities were part of the project. The thinking was that by encouraging urban sprawl this would make cities difficult to occupy by invaders - a great tactic for 19th century warfare built in the 20th century.) What would the US look like without the interstate highway project? It would probably have the great features of European cities technocrats have been complaining for years as being absent in the US - despite their best efforts to encourage them! (walkable city centers, viable mass transit, able to travel from place to place by modern high-speed rail)
5. Lockdowns: Were lockdowns really a strong move here? I'm sensitive to the argument that technocratic solutions are ideal for some problems, and it would seem like a pandemic such as this one would have been the best place to demonstrate that. Yet instead of good solutions based on solid epidemiological evidence, the technocratic solution was lockdowns - to the exclusion of many other, better solutions.
I wish we could have racked up a win here by implementing a different strategy. What if, after it became clear by early May who the vulnerable populations were, we implemented a cocooning strategy advocated by me and a bunch of others? What if we'd given free grocery and restaurant delivery to elderly people for a few months - at the cost of a few million dollars - to allow them to isolate more effectively? We could have paid people who'd already been infected (and recovered) to provide other services and helped high-risk people not be so lonely. It would have been a great example of how the technocrats got it right, implemented the obvious solutions based on good scientific evidence, saved lives, and kept the economy from tanking.
That's not what happened, though. Because the problem with technocracy is that it has to contend with Public Choice Theory, and it rarely comes out the victor in that exchange.
An interesting companion article: In Praise of Passivity by Huemer: https://spot.colorado.edu/~huemer/papers/passivity.htm
As an economist, I've generally appreciated Weyl's criticisms of technocracy, but I ultimately don't find them extremely useful from an EA perspective. EA's are now mostly interested in ensuring the long-term future (10,000 + years) goes well and because future generations don't exist yet, it is hard to see how 'local knowledge' considerations could come into play.
In fact, I see this as part of the case for long-termism. Because future people can't vote or participate in the market, their preferences are for the most part ignored. If there was a way to incorporate these preferences into democracy and markets today, I would be very happy! But, I don't really see how that would be possible. Technocracy seems to be the only option.
Great piece
> In light of the ignorance of typical political leaders and members of the general public, we might be tempted by the idea of rule by experts, as in Plato’s Republic. Unfortunately, when it comes to descriptive social theory, even the experts’ knowledge is unimpressive, as demonstrated recently by the social psychologist Phillip Tetlock. Tetlock conducted a fifteen-year study in which he collected tens of thousands of predictions from hundreds of political experts concerning matters within their areas of expertise (for example, would the economy slide into recession, would the Soviet Union survive, who would win the next Presidential election, and so on). Tetlock’s finding, in brief, was that the best experts did only slightly better than chance at predicting outcomes. When asked to assign probabilities to their predictions, experts proved systematically overconfident; for example, events predicted with 100% confidence happened less than 80% of the time.
https://spot.colorado.edu/~huemer/papers/passivity.htm
I reached out to Weyl on Twitter, and he says he'd be open to an adversarial collaboration on this topic: https://twitter.com/glenweyl/status/1355202263380320264?s=20
Pretty surprising that neither Weyl's article nor your critique mention futarchy. I would expect this to be one of the first things brought up, given that:
-- you are both pretty strongly Robin-Hanson-adjacent
-- two very common and salient critiques of technocrats are that they either (a) have the wrong incentives to make good empirical judgments or (b) conflate empirical judgments and value judgments, fail to realize their training confers no advantage in making value judgments, and arrogantly shut out the common people from any voice in value judgments
-- so a system explicitly designed to incentivize technocrats to make good empirical judgments, while leaving value judgments in the hands of the common people, seems like a pretty obvious thing to consider when discussing how to deal with common failings of technocracy!
> All of this is too bad, because if Weyl was right, that would be a huge point in favor of rational methods. Imagine pointing to a community that did something wrong for years, then learned that thing was wrong, then admitted it and changed - and imagine thinking of it as a strike against that community's methods!
Doesn't this violate the conservation of expected evidence? Rationalism can't win both ways, so if this being true were a point for Rationalism, then it being false should be a point against it, Bayes-wise (the size of points may be different inverse to expectation, but it would still be against). Since I don't really see how it being false is a point against Rationalism, then it feels like that couldn't actually be a point for it. Is my Bayescraft off?
The things Weyl says about Rationalism and EA come across as confused.
e.g. "The effective altruism movement...seeks to maximize the efficacy with which charitable donations are directed using standard rationalist methods. It is a tight-knit community that strongly privileges rationalist approaches over all other forms of knowledge-making (such as from the humanities, continental philosophy, or humanistic social sciences) and tends to dismiss input not formulated in rationalist terms. The community also has a strong and explicitly stated view that its activities uniquely contribute to the achievement of 'the good': of their top five recommendations of most productive careers by a leading community organization, two suggest being a researcher or support staff within the movement, and two others recommend working on the AI alignment problem (see the next point)." This sounds mildly like "Beware, EAs are an elitist cult who doesn't listen to outsiders". Like much of the article, it's subtle jabbing couched in academic language in order to mask its intellectual vacuousness.
"Until recently, much of the analysis and funding emerging from the community has pointed towards a focus on extremely unlikely but potentially catastrophic risks, such as alien, asteroid or biological catastrophes." Granted it was only 3 years ago that I pledged to give 10+% of my income to charity for the rest of my life, but I missed the memo about Our Focus On Aliens And Asteroids.
"Yet, interestingly, the conclusions of the analysis emerging from the community increasingly undermine these foci and the approach of the community more broadly. In particular, recent research in the community suggests that the greatest and most probable risks to be avoided are anthropogenic" Um, hello? AI catastrophic risks *are* anthropogenic! So are the "biological catastrophes" we worry about! What, did you think we thought that a naturally-occuring virus was likely to wipe out humanity in the next 100 years?
"Leaders in the community have in turn suggested that the most effective ways to avoid these are likely finding solutions to problems of political organization and legitimacy of social systems to help reduce the likelihood of conflict or inability to cooperate in the provision of critical global public goods." Actually, yes. Very astute. Unfortunately, Weyl doesn't recognize that the ability of the community to change its mind is arguably its greatest strength. On Tim Urban's "Thinking Latter", this is the main difference beteween endless conflict between echo chambers, and productive cooperation and understanding.
"Ironically, those “political” goals, social reforms, and public confidence building are precisely the sorts of activities that effective altruists have long viewed as 'non-rigorous' and ineffectual." Oh really? Please quote me any well-known community member ever claiming that these things were "ineffectual"?
"Worse, the extremely elitist, segregated, and condescending approach to philanthropy encouraged by the community has created widespread public backlash..." Um, what backlash? I think if there's a backlash it'll be generated by people like yourself who are happy to portray it inaccurately, in much the same way as there was a "backlash" against having a "socialist" president who "wasn't born in America" and "backlash" against a "fascist" president held in check only by "The Resistance".
This is a well-written essay and it brings up very interesting arguments. I also wish I hadn't read it. I try to stay away from culture war topics as much as possible, because when they do end up finding me, they occupy my attention much more than they deserve.
A filter option would make me very happy, to keep receiving posts about the history of amphetamines, but not essays about culture war topics I otherwise would not have found. (A bit like https://xkcd.com/2071/). Yes, I'm aware I'm wishing for a stronger filter bubble, even though filter bubbles have been found to be the source of all modern evil.
Why is "I̶t̶'̶s̶ ̶n̶o̶t̶ ̶a̶ ̶s̶t̶o̶r̶y̶ ̶t̶h̶e̶ ̶J̶e̶d̶i̶ ̶w̶o̶u̶l̶d̶ ̶t̶e̶l̶l̶ ̶y̶o̶u̶" in a notably different (sans serif, for starters) font than the rest of the piece?
You mind find the recent fiasco with awarding UK A-level and GCSE grades last year an interesting case study:
:- It was not possible to hold exams this year, but we still had to award qualifications in some way, and this couldn't be postponed because of e.g. university places.
:- The government decision was that teachers should make, and announce, predictions of the grades their students would receive, and then an algorithm would be used comparing those grades against grades of previous years at the same institution (and possibly previous grades of the same students?) to normalise those.
:- In ordinary years teachers release predicted grades too, and those grades are consistently higher on average than the actual grades turn out to be.
:- When the algorithm awarded many students lower grades than the unprocessed predictions from their teachers, there was outcry. The government was forced to back down and a week or two later the original, unnormalised grades were awarded.
:-The outcry was exacerbated by the fact that for very small classes it was felt that there wasn't enough data to normalise, and so the raw, probably-too-high grades were used in those places only; this turned out to disproportionately benefit private schools with children of rich parents at them.
:- This has put universities in a terrible bind. Universities typically make initial offers to students early in the year - "we will definitely take you if you get two A's and a B" - and then, if not enough students make their demanded grade (which will be the case for a lot of universities each year) they take additional students once the grades are announced, through a process called clearing. This year, clearing happened /before/ the government backed down, all places were filled, and then it turned out that a bunch more students had also made their predicted grades, and places had to be found for them to, which will cause significant overcrowding problems for the next three or four years.
:- It's probably all going to happen again this year.
Having read through the original article, I do think Weyl has a point on (some) discussion of AI alignment, a lot of discussions really are, I quote: "a bit like posing the “genius dictator alignment problem”: how can we ensure that if there is a brilliant dictator, he will serve the interests of the broader public?"
I think the solution is going to involve limiting the power that computers have over people and ensuring they remain accountable (probably through the use of other AIs). Humans are the ones developing AI, this really is just the ancient question of how to limit the power some people have over other people.
Long term, I think I worry more about the human alignment problem than the AI alignment problem - which human values will we align AI to?
The topic of "making technocrats' plans legible to the populace" reminds me of a recent Matt Yglesias essay: https://www.slowboring.com/p/making-policy-for-a-low-trust-world.
The idea is that policy should "do what is says on the tin": make it easy for everyone, regardless of whether they agree with you, to understand what it is say you're going to do, and then to see whether or not you actually succeeded in doing that thing.
A simple policy that is less "optimal" or less fine-tuned may have better outcomes because its legibility makes it more likely to get the requisite political support. A meta-rational technocrat should be optimizing for policy legibility!
This one made my head hurt...and it is Friday.
I am someone who is guilty of what you are criticizing, definitely. I appreciate this essay tremendously and definitely agree. Too many people want to use "technocracy" or some equal and opposite thing (let's say "irrationality") as their One Move That Ends the Debate, when at the end of the day you have to get into the nitty gritty. As a humanist, I say, ah, there's no mechanism for avoiding that nitty gritty! As a chastened one, I add, I can't really find much fault in the nitty gritty you provide here, and also, I don't really think you or most notables in your community believe that mechanism must supplant human judgment everywhere and anywhere. At any rate, very glad to see you writing again so I can read essays like this one.
"In the Languedoc there is a vineyard that teaches us an important lesson about textbook learning and its application to the world. In the early Seventies it was bought by a wealthy couple, who consulted professors Emile Peynaud and Henri Enjalbert, the world’s leading academic oenologist and oenological geologist respectively. Between them these men convinced the couple that their new vineyard had a theoretically ideal microclimate for wine-making. When planted with theoretically ideal vines whose fruits would be processed in the optimal way according to the up-to-date science of oenology, this vineyard had the potential to produce wine to match the great first growths of Bordeaux. The received wisdom that great wine was the product of an inscrutable (and untransferable) tradition was quite mistaken, the professors said: it could be done with hard work and a fanatical attention to detail. The couple, who had no experience of wine-making but much faith in professorial expertise, took a deep breath and went ahead.
If life were reliably like novels, their experiment would have been a disaster. In fact Aimé and Véronique Guibert have met with a success so unsullied that it would make a stupefying novel [...] The first vintage they declared (in 1978) was described by Gault Millau as ‘Château Lafite du Languedoc’; others have been praised to the heights by the likes of Hugh Johnson and Robert Parker. The wine is now on the list at the Tour d’Argent and the 1986 vintage retails at the vineyard for £65 a bottle. The sole shadow on the lives of these millionaires is cast by the odd hailstorm."
https://paulseabright.com/wp-content/uploads/2012/01/Paul-Seabright-reviews-%E2%80%98Seeing-Like-a-State%E2%80%99-by-James-C.-Scott-%C2%B7-LRB-27-May-1999.pdf
Have you read "Skin in the Game"? Everything you spoke about could be looked at from localist and skin-in-the game perspective.
So I read his comments on EA and AI alignment.
Why even bother listening to this guy lol. Is the argument that if he's this wrong about EA and AI, he must be really good at other stuff?
Naively I'd guess the opposite, that induction is more likely than anti-induction here (as it is the case in most other domains).
Good post. Simultaneously encouraged and disappointed that the most controversial example in the comments is not the culture war-adjacent example, but instead whether strong mitigation was needed for COVID-19.
Please, stop using Brasília as an example of a failed technocratic project. Niemeyer didn't build a Modernist village, because he built no village at all - he merely designed some (very popular) public buildings and spaces. Brasilia (and a good deal of the Federal District) was projected by Lucio Costa. I lived there for 3 months, and most of the residents (specially family people) love it. And there was no bottom-up alternative, since there was nothing in its place. What really bothered me about the city is some peculiar governance issues (though I can't say other Brazilian capitals, like São Paulo antes Rio, fate better) and the artificial environment provided by it's status as the federal capital city.
The three examples you give; mandatory vaccination, desegregation and the federal highway system; prove the point you are trying to disprove. The heavy handed all or nothing approach that the federal government took in desegregating schools resulted in a devastating backlash, a backlash that is at the root of how the Republican party got coopted by extreme Libertarians (see Democracy in Chains). What should have been done is they needed to start with desegregating the teachers first.
When Japan and Europe were investing in trains, the US built freeways. What's happened? We have a world that revolves around the personal automobile with horrendous environmental consequences and the endless nightmare of congested roads and marathon commutes and we are stuck with it. Talk to anyone in their eighties. They all remember the freeways being shoved down their throats. The freeways ruined American cities. Gutted them.
As for mandatory vaccination, I don't actually know much about that only that the most basic human instinct is freedom. As soon as you try and control behavior or actions, you are asking for trouble. The really simple solution always was and will be: Wrong thing hard, right thing easy. Vaccination is free and efficient, if you don't want to be vaccinated, you have to pay a yearly fee. $500? That should do it. That would ensure so few people remain unvaccinated that it wouldn't be a concern. Better yet, give people $10 for getting vaccinated. You'd be surprised how little money you have to pay people to ensure their co-operation.
I think another benefit of simple models is that sometimes they're the only thing you can get through your political system.
By the time your simple policy has ground through the multifarious incentives of the political system it's likely kludged beyond imagining.
If you start with something complex it may never see the light of day.
A debate that seems to coming up often is whether farm collectivisation counts as a genuine technocratic act if it wasn't actually scientifically valid. The problem is that the authoritarian governments that implemented it would of course have claimed that their policies were scientific; except that all the scientists were ideologically compromised by the state and members of the public couldn't see the data. So perhaps one component of a true technocracy is open science - the studies motivating a policy decision should be, as far as possible, visible to and experimentally verifiable by the public.
This feels like the narcissism of small differences but it’s entirely possible that there’s a meaningful disagreement here I’m missing.
"Autarchy vs. democracy."
I'd recommend using "autocracy" instead. "Autarchy" is easily confused with "autarky," which is obviously not what you mean.
Maybe I missed it, but what's the original Weyl article that this post is a response to? Can someone post a link?
The "Republicans bad" invective was a little over the top in this one.
I haven't read this thread, so apologies if this is redundant, but many of the distinctions you mention, along with many additional examples, were discussed in The Future and its Enemies back in 1998. (BTW, Technocracy was actually a political movement in the 20th century, with the idea that all of society and economy should be centrally directed along the lines with which total war had been waged.)
In my wording, technocracy is an ideology fundamentally opposed to spontaneous orders that emerge from the interaction of "private" or "autonomous" choices by subsystem actors; those subsystem actors can be as rationalist and expert-driven as they choose, and the technocrat will still oppose their execution and (unplanned) interaction. Moreover it insists that virtue inheres in the mere intention to predict and control the output of the total system, and whether the effort is successful according to some objective function is of secondary importance. There are technocrats who want to subject system planning to majoritarian control, those who want it delegated to a coterie of experts of some sort, etc.
But it is a category error IMO to apply the term "technocracy" to something operating at a subsystem level--say, a college admissions office, or the quality-control department of a factory--embedded in a larger competitive and cooperative set of entities whose actions it must take as given. Designing an entire city from scratch, or subsuming an entire agricultural sector into collective farming, sure. The "-cracy" aspect should be taken seriously as referring to political control over a whole society, not to the private endeavors of actors within the system.
"The Soviet leadership wasn't "meritocratically chosen". "
What is 'meritocracy' Scott, and what evidence do you have that they weren't chosen in a meritocratic manner? Similarly, is the supposed 'failure' of Soviet collective farming something you actually studied using objective agricultural data, or is it just a vague feeling you have?
You keep using the word "compact" where you probably meant "convex".
God I missed your writings. Good to have you back!
Funny. I was just reading about how Taiwan keeps discussions on an online policy forum productive, by allowing likes but instead banning replies.
I thought that was an interesting contrast to the comment moderation discussion last week, where SSC prefers to ban likes and encourage replies.
The Taiwan system also includes some other features I would have called "technocratic," to nudge people towards civil, productive discussion.
So I come on to mention it, only to find it already under discussion as an example contra technocracy. Highly unexpected!
Given my confusion, I agree there might be a gap in definitions here.
I still think it's worth reading about that system on its own merits, completely separate from the way it's being referenced in this discussion. One can appreciate developments in Taiwan without taking a firm stance on whatever either side is ultimately calling "technocracy."
https://www.technologyreview.com/2018/08/21/240284/the-simple-but-ingenious-system-taiwan-uses-to-crowdsource-its-laws/
Algorithmic redistricting is the only viable approach. Humans are absolutely unfit to draw districts. The evidence here is just profound.
https://www.rangevoting.org/TheorDistrict#Paradigm
Without commenting either way on the broader argument here, just wanted to point out that some of your examples in the beginning are odd. How is school desegregation an example of technocratic rule? It has been the result of intense grassroots pressure from the civil rights movement, preceded by a civil war. Stranger still to describe it as an attempt to get schools to more closely adhere to the Constitution, which was written by slave owners who couldn't have imagined such a thing. Also fail to see how scientists sounding the alarm about global warming are "technocrats". The scientific community hasn't called the shots in the (failed) struggle against climate change, and our technocratic rulers have miserably failed us.
Wow this post was incredible. A couple of years back somebody recommended slate star Codex to me and I read a few posts and determined that Scott was just a horrible writer with no ability to structure or organize his thoughts in a clear concise way. but here he has done that so masterfully that I have to assume I just somehow misjudged him before.
I was actually a co-founder of a nonprofit that advocates for a political reform called approval voting, and it got its legs after many years of struggle due to a grant from the open philanthropy foundation led by the effect of altruism movements William McCaskill. one of the other founders of that non-profit is a Princeton math PhD who created a redistricting algorithm called the shortest split line algorithm. I wonder if that's one of the ones Scott is thinking about in this post.
On a bit of a tangent, that nonprofit recently brought approval voting to St Louis, the second US city to use it after starting in Fargo.
https://electionscience.org/press-releases/diverse-coalition-brings-approval-voting-nonpartisan-election-reforms-to-st-louis/
Because of this post I went back and reread the Seeing Like a State review. The review was written in 2017 (around when I read the book myself); in 2019 I got big into Roman history for personal reasons (dorkiness).
And I gotta say, it feels like the big thing missing from this conversation is Imperial Rome. The Romans loved big round numbers and big tiled rectangles. Their fundamental statecraft innovation was taking a bunch of people from deeply embedded rich local cultures, dragging them thousands of miles away, making them march around in rectangles, build perfectly rectangular forts, and then eventually mustering them out to perfectly rectangular farm allotments in unrelated parts of the empire—farms so rectangular you can still see the rectangles etched across the face of Europe today.
And it worked LIKE CRAZY. Through the magic of rectangles, Rome took itself from a relatively minor city state marginal to both the Greek and Punic worlds to the predominant social order of the Mediterranean basin. And, as far as I can tell—though I'm far from a scholar—Rome mostly didn't experience the failure modes associated with, e.g., Soviet collective farms or Prussian scientific forests. I don't hear, for example, that the veteran farms were plagued with collapse or underproduction.
Now, of course, RECTANGLES(TM) are not the whole story. As with many successful empires Rome relied on all sorts of rich syncretism of assimilated and affiliated cultures. Auxiliary forces fought in the armies; foreign gods were smushed willy-nilly into the pantheon; and so on. So it's certainly not a pure triumph of High Modernism two millennia early.
But maybe this puts a bit of a different spin on the Love of Rectangles and its sundry failures. The fact of the matter is that Rectangle Magic, properly applied, is extremely powerful—not just in theory but in practice. And not every use of Rectangle Magic yields Rectangle Rot. And to get the gains out of Rectangle Magic you kind of have to bet big on it and commit in a big way—you get the legions out of a bunch of technocratic reforms and top-down dictates.
So maybe the High Modernists were kinda . . . on to something? Or rather, they were investing in a strategy that had paid out before: take some domain of human life which is currently a bit of a mess and ADD RECTANGLES UNTIL IT CRIES OUT FOR MERCY. Which has a track record other than failure!
Makes me kind of wish that there was a companion book to Seeing Like a State about the record of success and growth associated with rationalization, from the state's perspective. You could call it Seeing Like a State, maybe.
"This isn't just true for colleges - we know that giving everyone IQ tests and letting the top scorers into gifted programs ends up with better representation of gifted minorities than letting teachers use their judgment."
This isn't correct, see the paper:
"We test this hypothesis using data from a unique natural experiment conducted by a large and diverse school district in the state of Florida (hereafter “the District”). State law dictates that students must achieve a minimum of 130 points on a standard IQ test to qualify for gifted status. English language learners (ELLs) and free-or-reduced price lunch (FRL) participants are subject to a lower 116 point threshold, known as “Plan B” eligibility."
Plenty of my highly educated, upper middle class liberal friends would say they openly support technocracy. We must run in different circles.
> No one ever defends technocracy
Well, IIUC, the term originates with this (pro-technocracy) political movement:
https://en.wikipedia.org/wiki/Technocracy_movement
> No one ever defends technocracy
Well, IIUC, the term originates with this (pro-technocracy) political movement:
https://en.wikipedia.org/wiki/Technocracy_movement
What does "metis" mean?
"Climatologists created complicated formal models to determine how quickly global temperatures might rise"
I apologies because this is about climate modelling, and nitpicking your off-the-cuff example, but I thought I saw an opportunity to say something useful so here goes:
So there is a big and important misconception here: climate models are NOT currently (properly considered) part of the scientific case for global warming; and people saying they are, convinces a lot of climate-change-skeptics that they are on a firmer footing than they actually are -- a lot of climate-change-skepticism is pointing out limitations in model reliability. And their prominence in the news may fuel climate-skepticism (when people generally learn to treat similar concepts like economic-models and forecasts with friendly contempt).
But this is irrelevant because the scientific case around global warning stands independent of modelling.
The go-to source for evaluations of the effectiveness of state-of-the-art climate-models is the IPCC (the UN body tasked for compiling evidence on climate change) regular reports on "The Physical Science Basis", each of which contains a section "Evaluation of Climate Models" (or some paraphrase). They're all available online and very readable. Go check out.
The 4th edition of their report (a few years back) devotes 74 pages to this topic. The newest report is a short refresher on this pointing out which things have progressed since then.
I guess one way of looking at it is that the "aggregate climate model" the IPCC uses, is the aggregate of predictions of many different models, many of which have completely different assumptions about dominant mechanisms. They use this aggregation because some models are good at some things and some good at other things. But I think it's difficult to build a scientific case on an aggregate of contradictory scientific cases?... not like "well, if you believe X, then Y, and if you believe not-X, then also Y, so Y" -- that's a fine argument -- but this argument would be "I think X and I think not-X, and their aggregate is Y". They are a best effort at prediction. They aren't a case. The rest of the 1007 page report (AR4) is the case.
I have a bone to pick with all of your examples. I think they are all interesting and subtle and not so clearcut. The interstate highway system is clearly a good thing and marvel of engineering, but many beautiful neighborhoods were paved, and it ended up creating the problems Not Just Bikes is always complaining about. As for school desegregation, I don't know much about this, but didn't busing and integration also come with seriouses increases in crime? Mandatory vaccination seems like a good thing for smallpox, but there is a serious debate about COVID vaccine mandates, given vaccine side effects. As for climate change, I am almost certain that any climate legislation will do much more harm than good. The climate movement is completely deranged, anti-nuclear, calling for a radical transformation that if seriously pursued would be disastrous. And the COVID lockdowns were according to Bhattacharya the greatest public health disaster in history.
History is written by elites, so whatever they do (and did) is wise and good and certainly not a big mistake.