Spontaneous order, local knowledge and other castings of the relationship of local inscrutability and totalising modernity in the social order. I have no original thoughts on this, but I like to keep links on this theme where I can see them so that they don’t bite me.
TODO: find Jaron Lanier quote about APIs, and legibility. Muse about standards.
The rationality of the Great Society
Scott Alexander, Contra Weyl On Technocracy
🏗, quote Constantin, In defense of individualist culture, and Hayek’s “constructivist fallacy”, Timothy Morton’s Hyperobjects, and Berkes and Folke’s “local knowledge”, pragmatist notions of a belief’s “cash value”, local versus global truth, and all the other dissections of these problems, and wonder about idiosyncratic spontaneous group order etc. Discuss Social Capital and other economic framings as a method for making metis “legible”. The Master Currency displacing other possible currencies. Or, to have this phrased in a manner intelligible to management, Florent Crivello, The Efficiency-Destroying magic of tidying up. Contrast this with the Robin Hanson opinion on the grab-for-power that invoking metis can mask:
Apparently most for-profit firms could make substantially more profits if only they’d use simple decision theory to analyze key decisions. Execs’ usual excuse is that key parameters are unmeasurable, but Hubbard argues convincingly that this is just not true.[…]
I say that their motives are more political: execs and their allies gain more by using other more flexible decision making frameworks for key decisions, frameworks with more wiggle room to help them justify whatever decision happens to favor them politically. Decision theory, in contrast, threatens to more strongly recommend a particular hard-to-predict decision in each case. As execs gain when the orgs under them are more efficient, they don’t mind decision theory being used down there. But they don’t want it up at their level and above, for decisions that say if they and their allies win or lose.
Sam Popowich’s invective Lawful Neutral re-implements some of these analyses, with “liberalism” assuming the role of the state in his version:
Liberalism — like the necessarily undemocratic capitalism for which it serves as an alibi — seeks to reduce human life to a predictable, exploitable, profitable minimum. It adopts the simplest social ontology (individualism) to make its formalism work, to make its algorithms or procedures appear universally applicable. Liberalism demands a strict division between form (the system of rules) and content (the messy details of social entanglement) to make reality tractable to its logic. Artificial intelligence likewise requires that its form (code) be separate from its content (data). Machines require the simplest possible data on which to work (binary numbers, for example) to make their procedures uniform and generalizable and the world computable.
He also equates the project of procedural AI with liberalism.
If AI can’t necessarily replicate human intelligence, it nevertheless precisely models the sort of intelligence needed to make liberalism coherent.
From there, he argues, the system of liberalism produces individualism and negates community, which is antithetical to the messy social world of humans otherwise. Not covered in his article: Mass systems of governance which do not have the asserted pathologies of liberalism and yet also produce great good for a great number. (I imagine he would suggest Marxism? Some kind of anarchism? Libertarian something? Neofeudalism?)
C&C Robin Hanson’s imagined dialogue about the role of simplification in models in science and engineering, and the application of reductionism to the study of people. (tl;dr fraught, but the world is too complex to interact with, if you do not choose some simplification with which to model it, so choose simplifications based on opportunity costs.)
Antidote, M. John Harrison
I’m not interested in an embodied and localised knowledge. I had enough of it as a child in the early 1950s, among people whose top argument was, “Because I know better.” They didn’t want the NHS. They didn’t want vaccination. They didn’t want the kids that survived to waste their time on education. They didn’t want science. All sense was common sense: they were the well, and your role as a child was to drink what you were given. Anything else, from the welfare state to astrophysics, was a challenge to traditional hierarchies. After you’d tried, and had it drilled into you how worthless your fancy new ideas were, your ambition was to quickly and quietly exit their radius of control and enter the de-localised intellectual funfair of modernity, with its fantastically advanced concepts such as “abstract thought”. Your secondary ambition was to work on a politics that got shot of all that forever. That’s really what the 60s was about if you lived where I lived. It was a revolt about what kind of knowledge you could have of the world, and how you could get it. I never regretted running away from the Trump-like epistemics of postwar semi-industrialised, semi-rural England. All I regret is that we didn’t quite achieve full escape velocity and get rid of its limiting ideas forever, so they couldn’t crawl back and infect everything again. I never want to return to that nightmare, or sympathise with it, or “understand” it, or give it any more than this single paragraph of the oxygen of analysis.
Policy and Statistical learning
TODO. Brief digression on how legibility and management looks as a statistical learning problem. We know that constructing policies is costly in data, and we know that administrative procedures frequently do not have much data from repeated trials of what works. We also know that coming up with policies (in a machine learning or in a political definition) is computationally challenging and data hungry. How does the need to bow to the ill-fitting bureaucracy of the Great Society resemble having to work with an underfit estimator of the optimal policy? What does that tell us about, e.g. optimal jurispudence? Possibly something. Or possibly the metaphor doesn’t work; after all, what is the optimisation problem one solves?