31/12/2016

Weapons of Math Destruction by Cathy O'Neil


Weapons of Math Destruction (2016) by Cathy O'Neil

Solid work of practical ethics, covering the rights and wrongs of applied statistics in general, but particularly our mass, covert automation of business logic in schools, universities, policing, sentencing, recruitment, health, voting... Much to admire: she is a quant (an expert high-stakes modeller) herself, understands the deep potential of modelling, and prefaces her negative examples with examples of great models, methods of math construction - the moneyball wonks, the FICO men,
in mid-2011, when Occupy Wall Street sprang to life in Lower Manhattan, I saw that we had work to do among the broader public. Thousands had gathered to demand economic justice and accountability. And yet when I heard interviews with the Occupiers, they often seemed ignorant of basic issues related to finance...
They are lucky to have her!

A 'Weapon of Math Destruction' is a model which is unaccountably damaging to many people's lives. She doesn't cash out her criteria ("Opacity, , but here's an enumeration.

Standardisation - the agreement of units and interfaces that everyone can use - is one of those boring but absolutely vital and inspiring human achievements. WMDs are a dark standardisation.

In her account, a model doesn't have to be predatory to be a WMD: for instance the recidivism estimator was first formulated by a public-spirited gent.



She says "efficiency" when she often means "accuracy, and thereby efficiency". This makes sense rhetorically, because she doesn't want to give the predatory models a halo effect from the less backhanded word "accurate". Also lards the texts with "Big Data"

The writer Charles Stross, asked recently to write the shortest possible horror story, responded: "Permanent Transferrable Employee Record". O'Neil makes us consider that this could be easily augmented with "including employee medical histories".

She covers most of the objections I would make to an unsubtle author:
It is true, as data boosters are quick to point out, that the human brain runs internal models of its own, and they’re often tinged with prejudice or self-interest. So its outputs—in this case, teacher evaluations—must also be audited for fairness. And these audits have to be carefully designed and tested by human beings, and afterward automated. In the meantime, mathematicians can get to work on devising models to help teachers measure their own effectiveness and improve.


Optimising for truth. Wants fairness instead. Can you be fair without knowing the truth?
Not optimising for profit, that is bad, though one argument of the profit function is truth.

She does not include inaccuracy as a named criterion for WMDs, but her discussions sometimes require it. This is maybe the core shortcoming of the book: it doesn't wrestle much with the hard tradeoff involved in when modelling unfair situations, e.g. living in a bad neighbourhood which increases your risks and insurance costs through no fault of your own. She comes down straightforwardly on the direct "make the model pretend it isn't there" diktat.

But then she notes a case where fairness trumping: value-added teacher rating:
The teacher scores derived from the tests measured nothing. This may sound like hyperbole. After all, kids took tests, and those scores contributed to Clifford’s. That much is true. But Clifford’s scores, both his humiliating 6 and his chest-thumping 96, were based almost entirely on approximations that were so weak they were essentially random.

The problem was that the administrators lost track of accuracy in their quest to be fair. They understood that it wasn’t right for teachers in rich schools to get too much credit when the sons and daughters of doctors and lawyers marched off toward elite universities. Nor should teachers in poor districts be held to the same standards of achievement. We cannot expect them to perform 1miracles.

So instead of measuring teachers on an absolute scale, they tried to adjust for social inequalities in the model. Instead of comparing Tim Clifford’s students to others in different neighborhoods, they would compare them with forecast models of themselves. The students each had a predicted score. If they surpassed this prediction, the teacher got the credit. If they came up short, the teacher got the blame. If that sounds primitive to you, believe me, it is.


My preferred measure would be to not prevent models from being rational, but instead make transfers to the victims of empirically unfair situation. (This looks pointlessly indirect, but price theory, and the harms of messing with them, is one of the few replicated economic.) My measure has the advantage of not requiring a massive interpretative creep of regulation: you just see what the models do as black boxes and then levy justice taxes after.
Statistically speaking, in these attempts to free the tests from class and color, the administrators moved from a primary to a secondary model. Instead of basing scores on direct measurement of the students, they based them on the so-called error term — the gap between results and expectations. Mathematically, this is a much sketchier proposition. Since the expectations themselves are derived from statistics, these amount to guesses on top of guesses. The result is a model with loads of random results, what statisticians call “noise.”
Past unfairness gives the prior. Marx's critique of emp.
Innocent people surrounded by criminals get treated badly, and criminals surrounded by a law-abiding public get a pass. And because of the strong correlation between poverty and reported crime, the poor continue to get caught up in these digital dragnets. The rest of us barely have to think about them.
She flip-flops between thinking that false positives are the problem, and that any positives based on uncomfortable variables are the problem. Surprisingly big evidential gaps:
By 2009, it was clear that the lessons of the market collapse had brought no new direction to the world of finance and had instilled no new values. The lobbyists succeeded, for the most part, and the game remained the same: to rope in dumb money. Except for a few regulations that added a few hoops to jump through, life went on.


Engrossing car crash of American higher education. She credits the US News ranking with creating the whole mess though, which can't be right. Certainly it had some effect, and could fix some of its harm by including tuition fee size as a negative factor.

even those who claw their way into a top college lose out. If you think about it, the college admissions game, while lucrative for some, has virtually no educational value. The complex and fraught production simply re-sorts and reranks the very same pool of eighteen-year-old kids in newfangled ways. They don’t master important skills by jumping through many more hoops or writing meticulously targeted college essays under the watchful eye of professional tutors. Others scrounge online for cut-rate versions of those tutors. All of them, from the rich to the working class, are simply being trained to fit into an enormous machine—to satisfy a WMD. And at the end of the ordeal, many of them will be saddled with debt that will take decades to pay off. They’re pawns in an arms race, and it’s a particularly nasty one.


Predictive models are, increasingly, the tools we will be relying on to run our institutions, deploy our resources, and manage our lives. But as I’ve tried to show throughout this book, these models are constructed not just from data but from the choices we make about which data to pay attention to—and which to leave out. Those choices are not just about logistics, profits, and efficiency. They are fundamentally moral...

Big Data processes codify the past. They do not invent the future. Doing that requires moral imagination, and that’s something only humans can provide. We have to explicitly embed better values into our algorithms, creating models that follow our ethical lead.

Quick, sharp-toothed and filling. Extra half a point for requiring no technical background; it proves clear thought
4/5.


****************************************************************************************

Other highlighted passages:
Anywhere you find the combination of great need and ignorance, you’ll likely see predatory ads.

All of these data points were proxies. In his search for financial responsibility, the banker could have dispassionately studied the numbers (as some exemplary bankers no doubt did). But instead he drew correlations to race, religion, and family connections. In doing so, he avoided scrutinizing the borrower as an individual and instead placed him in a group of people — what statisticians today would call a “bucket.” “People like you,” he decided, could or could not be trusted.

Fair and Isaac’s great advance was to ditch the proxies in favor of the relevant financial data, like past behavior with respect to paying bills. They focused their analysis on the individual in question—and not on other people with similar attributes. E-scores, by contrast, march us back in time. They analyze the individual through a veritable blizzard of proxies. In a few milliseconds, they carry out thousands of “people like you” calculations. And if enough of these “similar” people turn out to be deadbeats or, worse, criminals, that individual will be treated accordingly.

From time to time, people ask me how to teach ethics to a class of data scientists. I usually begin with a discussion of how to build an e-score model and ask them whether it makes sense to use “race” as an input in the model. They inevitably respond that such a question would be unfair and probably illegal. The next question is whether to use “zip code.” This seems fair enough, at first. But it doesn’t take long for the students to see that they are codifying past injustices into their model. When they include an attribute such as “zip code,” they are expressing the opinion that the history of human behavior in that patch of real estate should determine, at least in part, what kind of loan a person who lives there should get.

In other words, the modelers for e-scores have to make do with trying to answer the question “How have people like you behaved in the past?” when ideally they would ask, “How have you behaved in the past?” The difference between these two questions is vast. Imagine if a highly motivated and responsible person with modest immigrant beginnings is trying to start a business and needs to rely on such a system for early investment. Who would take a chance on such a person? Probably not a model trained on such demographic and behavioral data.

should note that in the statistical universe proxies inhabit, they often work. More times than not, birds of a feather do fly together. Rich people buy cruises and BMWs. All too often, poor people need a payday loan. And since these statistical models appear to work much of the time, efficiency rises and profits surge. Investors double down on scientific systems that can place thousands of people into what appear to be the correct buckets. It’s the triumph of Big Data.
Microcosm:
This is not to say that personnel departments across America are intentionally building a poverty trap, much less a racist one. They no doubt believe that credit reports hold relevant facts that help them make important decisions. After all, “The more data, the better” is the guiding principle of the Information Age. Yet in the name of fairness, some of this data should remain uncrunched.


**************************************************************************

Insurance is an industry that draws on the majority of the community to respond to the needs of an unfortunate minority. In the villages we lived in centuries ago, families, religious groups, and neighbors helped look after each other when fire, accident, or illness struck. In the market economy, we outsource this care to insurance companies, which keep a portion of the money for themselves and call it profit.


Mistaken in the US: the "loss ratio" of US insurance as a whole is greater than 100: loss-making except for financial return on held premiums.

**************************************************************************

Campaigns scoring voters for ad targeting is new to me:
The campaigns use similar analysis to identify potential donors and to optimize each one. Here it gets complicated, because many of the donors themselves are carrying out their own calculations. They want the biggest bang for their buck. They know that if they immediately hand over the maximum contribution the campaign will view them as “fully tapped” and therefore irrelevant. But refusing to give any money will also render them irrelevant. So many give a drip-feed of money based on whether the messages they hear are ones they agree with. For them, managing a politician is like training a dog with treats. This training effect is all the more powerful for contributors to Super PACS, which do not limit political contributions.

The campaigns, of course, are well aware of this tactic. With microtargeting, they can send each of those donors the information most likely to pry more dollars from their bank accounts. And these messages will vary from one donor to the next.
**************************************************************************

Occasional data-poor hyperbole:
Workers often don’t have a clue about when they’ll be called to work. They are summoned by an arbitrary program. Scheduling software also creates a poisonous feedback loop. Consider Jannette Navarro. Her haphazard scheduling made it impossible for her to return to school, which dampened her employment prospects and kept her in the oversupplied pool of low-wage workers. The long and irregular hours also make it hard for workers to organize or to protest for better conditions. Instead, they face heightened anxiety and sleep deprivation, which causes dramatic mood swings and is responsible for an estimated 13 percent of highway deaths. Worse yet, since the software is designed to save companies money, it often limits workers’ hours to fewer than thirty per week, so that they are not eligible for company health insurance. And with their chaotic schedules, most find it impossible to make time for a second job. It’s almost as if the software were designed expressly to punish low-wage workers and to keep them down.

The solution for the statisticians at St. George’s—and for those in other industries—would be to build a digital version of a blind audition eliminating proxies such as geography, gender, race, or name to focus only on data relevant to medical education. The key is to analyze the skills each candidate brings to the school, not to judge him or her by comparison with people who seem similar...

we’ve seen time and again that mathematical models can sift through data to locate people who are likely to face great challenges, whether from crime, poverty, or education. It’s up to society whether to use that intelligence to reject and punish them—or to reach out to them with the resources they need. We can use the scale and efficiency that make WMDs so pernicious in order to help people. It all depends on the objective we choose.
But how do know we what's relevant to medical education, except by correlation discovery?

It would also be a cinch to pump up the income numbers for graduates. All colleges would have to do is shrink their liberal arts programs, and get rid of education departments and social work departments while they’re at it, since teachers and social workers make less money than engineers, chemists, and computer scientists. But they’re no less valuable to society.


28/12/2016

what I said to you in 2016


I reviewed an old book about new China. I reviewed Use of Weapons, conceptual art, and Biblical literalism. I wrote a long poem about technical wonder. I poked about Civ V's ideology. I riffed on scientific and non-scientific knowledge. I listed a range of new and/or rare words. I was impressed by nonfiction about nonreal things.

I listed some free web services for developers. I gave an algorithm for working out why your boot drive doesn't work. I talked about songs about music critics and we listened to them. I reviewed British villainy, Accelerando, and an Estonian poet. I reviewed an excellent bedroom synthpop album. I listed more new and/or rare words with no clear theme. I reviewed books on social physics, 'The Hitch', and Japanese poetry. I reviewed the gentlest introduction to Bayesianism. I got very maternal about the former Bishop of Edinburgh. I listed more new and/or rare words, on maths and wrestling.

I applauded preppers with one hand. I listed more new and/or rare words, on science and the internet. I listed real magic spells. I echoed another's horror about new Russia. I tried to explain why Bukowski is not shit and bad. I was interviewed about nonstandard animal ethics. I listed various ways to save time, which is to save lives. I listed foreign words in some sense. I counted up the Black Panthers' murders and made you call them racist. I got very animated about a pop-punk song. I realised that machine learning is the answer to the Wittgensteinian challenge to conceptual analysis. I wrote a typology of automation and made a bunch of sci-fi economic policy proposals but failed to really evaluate them. I mocked my new profession but also srsly help what I do is ill-defined. I marvelled at the coarse dialect of English spoken when lots of money is implicated. I reviewed columns by the greatest scientific educator in the country. I reviewed a sexy moralist, an angry reasoner, and Martha Gellhorn who was both. I gave another name to the great divide among do-gooders.

I listed more new and/or rare words, on netspeak and Christian aliens. I reviewed a book about recent do-gooders, which worries that their strength is weakness but ends up thinking it isn't. I listed words which don't look technical but which are. I ranted about corporate self-help. I got very excited about someone saying obvious sensible things. I talked about arbitrary agent liberation, not totally seriously. I listed more new and/or rare words, with no theme at all. I investigated various buzzwords which have played the cultural role that "data science" currently plays. I reviewed a large gathering of effective altruists. I listed more new and/or rare words, on statistical dragons and channish. I modelled early modern science. I reviewed and codified Cathy O'Neil's model of bad models. I reviewed Joel on software, white nerds on rap, and Bennett on fame.



Previously: 2015, 2014, 2013.


Been reading, Q4 2016


(c) Uno Due Tre Fuoco #3 (2012) by Ekaterina Panikanova

This page is related to that page. You're reading something constructed using a rhetorical practice, something informed both directly and indirectly by the entire history of composition up until this point, from the Sophists to Derrida. But you're navigating it using pure logical statements, using spans of text or images that, when clicked or selected, get other files and display them on your screen. The text is based in the rhetorical tradition; the links are based in the logical tradition; and somewhere in there is something worth figuring out.

...the entire history of Western pedagogy [is] an oscillation between these two traditions, between the tradition of rhetoric as a means for obtaining power — language as just a collection of interconnected signifiers co-relating, without a grounding in "truth," and the tradition of seeking truth, of searching for a fundamental, logical underpinning for the universe, using ideas like the platonic solids or Boolean logic, or tools like expert systems and particle accelerators ... what is the relationship between narratives and logic? What is
sprezzatura for the web? Hell if I know. My way of figuring it all out is to build the system and write inside it, because I'm too dense to work out theories.

Paul Ford


When the data-driven approach... did not lead to immediate success — and occasionally even when it did — it was open to attack in a way that the old approach to decision-making was not ...whatever it is in the human psyche — this hunger for an expert who knows things with certainty, even when certainty is not possible — has a talent for hanging around. It’s like a movie monster that’s meant to have been killed but is somehow always alive for the final act.

– Michael Lewis




As well as my usual durability scores, I added in how each book is trying to affect you, using Julia Galef's types of books. Her model is that a book can offer you new Data, Theories to explain data, arguments for Values, or entire Thinking styles. She also assigns a number 1-5, roughly "Concreteness -> Generality". I should like Data books better than I apparently do.

OCT


  • So You've Been Publicly Shamed (2015) by Jon Ronson. Actually important. What angry people are doing to jokers and liars and fools, generally on political grounds. There is only one representative of the online shamers her (besides Ronson, who is reformed). You realise quickly that she is not especially hateful: she's just dim – she still thinks shaming is great, even after suffering it horribly and losing her job as a result of her own aggressive humourlessness and insensitivity. In her interview with Ronson, she shows no signs of empathy or learning. It is a tragic example of how addling identity can be.

    Contains one essential passage, the payload inamongst Ronson's ordinariness and self-deprecation: a human-rights lawyer pointing out the emotional power of noncriminal acts:
    “Let me ask you three questions,” he said. “And then you’ll see it my way. Question One: What’s the worst thing that you have ever done to someone? It’s okay. You don’t have to confess it out loud. Question Two: What’s the worst criminal act that has ever been committed against you? Question Three: Which of the two was the most damaging for the victim?”

    The worst criminal act that has ever been committed against me was burglary. How damaging was it? Hardly damaging at all. I felt theoretically violated at the idea of a stranger wandering through my house. But I got the insurance money. I was mugged one time. I was eighteen. The man who mugged me was an alcoholic. He saw me coming out of a supermarket. “Give me your alcohol,” he yelled. He punched me in the face, grabbed my groceries, and ran away. There wasn’t any alcohol in my bag. I was upset for a few weeks, but it passed.

    And what was the worst thing I had ever done to someone? It was a terrible thing. It was devastating for them. It wasn’t against the law.

    Clive’s point was that the criminal justice system is supposed to repair harm, but most prisoners — young, black — have been incarcerated for acts far less emotionally damaging than the injuries we noncriminals perpetrate upon one another all the time — bad husbands, bad wives, ruthless bosses, bullies, bankers.
    (It has been claimed that this phase of internet social justice is on its way out - that the tactic is now to "call in"- that is, to correct an offender, but also to appeal to the offender's humanity, to try to bridge the gap. We can hope this will gain traction (3 years and counting...). In the meantime a roaring subculture has been founded upon the glorification of bad behaviour and utterly unpersuasive flames.)

    Ronson investigates the possible solutions to finding yourself shamed: you can 1) refuse to feel bad (or at least refuse to show them you're bleeding), own the thing they're trying to shame you for, like Max Mosley. This only works sometimes. 2) You can hide from the internet, try to SEO the affair down to Google page 3, where no-one goes, like someone it would be counterproductive to name. 3) You can start over, asking for forgiveness like Jonah Lehrer. (There is none; the internet is not interested in you improving your behaviour.)
    4/5.
    [Library]
    [Theory #3, Values #1]



  • The Best Software Writing I (2003) ed. Joel Spolsky. Odd beast: a time capsule where half the items are of purely historical interest, and half are general and extremely wise arguments that are still not acted upon today. He had planned them to be annual collections, but they didn't happen, so this looks to represent more than one year's best. Recent enough to tell us something about the internet, though with lots of anachronism. But it's more at the lexical level - "weblog", "Sociable media" - than the semantic.

    Found (the eminent media researcher) danah boyd excessive and insulting: calling developers autistic, and people with several online accounts as multiple-personality disordered ("a person is one person. So all their activites have to be one person!") . Disappointing, typical social theory. She aggressively pushes a risky single-sign-in for all sites based on these polemics and nothing else.

    Contains helpful principles which will not age:, e.g. "if you can't understand the spec for a new technology, don't worry: nobody else will understand it either, and the technology won't be that important".
    4/5 [Various]



  • Strangers Drowning by Larissa MacFarquhar. Engrossing, inspiring, deep. Full review here.
    (Wasn't an index in my Penguin copy; perhaps a sign of the ebook's dominance, since an index was always but a shadow of full-text search - or perhaps just another mild technical challenge on our way to having no boring tasks to do.)
    5?/5.
    [Library]
    [Data #2, Values #2]



  • Doing Data Science (2014) by Cathy O'Neil and Kathryn Schutt. The first third is: Talking About Data Science. Two careful, socially conscious techies talking is nice, and you would never get the dozens of handy heuristics in this from a usual STEM textbook. Crunchier than it looks - half the value is in the dull-looking, unannotated code samples at the end of each chapter, and isn't spelled out. Pedagogy!

    It is galling, then, that the data for chapters 6 and 8 has already link-rotted away. And half of the cool startups who came to talk to the class are dead and forgotten.

    Highly recommended for outsiders and newsiders.
    4/5.
    [Thinking #1, Theory 5 #2]



  • Forward Book of Poetry 2017 (2016) by Various. Mostly bad. I adore Harry Giles' verse; his big one, 'Brave', is a roaring, bouncing Orlando Furioso schtick with more point and verbal invention than the rest summed up, paist-apocalptic RPGs and all. The eventual winner, Tiphanie Yanique, is particularly glib: a series glorifying insensitivity.

    3/5 but see down the bottom of this post for 4*/5.
    [Library]
    [Data #1, Values #3]



  • The Codeless Code (2010) by "Qiless Qi". Parables about software development, violent and twee. Overwrought, and you can get them in a minute or two each time, unlike the bizarre originals which demand convoluted confabulation.

    It is a passable instance of an important genre: the self-conscious romanticisation and storification of highly abstract and highly novel things. We need such things; otherwise those of us without internal wellsprings of meaning will find it boring, and will thus never excel; otherwise a culture will never grow, and nothing human lasts without growing a culture.

    "Ah!", you say, "But Yudkowsky did just this, and got roundly mocked and called a cult leader and divers other bad things." Yes: that is the main tax we pay to be on the internet. I think of Yudkowsky as George Eliot thinks of Carlyle (though she hated him btw):
    ...the highest aim in education is analogous to the highest aim in mathematics, namely, to obtain not results but powers, not particular solutions, but the means by which endless solutions may be wrought. He is the most effective educator who aims less at perfecting specific acquirements than at producing that mental condition which renders acquirements easy, and leads to their useful application...

    On the same ground it may be said that the most effective writer is... he who rouses in others the activities that must issue in discovery, who awakes men from their indifference to the right and the wrong, who nerves their energies to seek for the truth and live up to it at whatever cost... he clears away the film from your eyes that you may search for data to some purpose. He does not, perhaps, convince you, but he strikes you, undeceives you, animates you. You are not directly fed by his books, but you are braced as by a walk up to an alpine summit, and yet subdued to calm and reverence as by the sublime things to be seen from that summit. Such a writer is Thomas Carlyle.

    It is an idle question to ask whether his books will be read a century hence: if they were all burnt as the grandest of Suttees on his funeral pile, it would be only like cutting down an oak after its acorns have sown a forest. For there is hardly a superior or active mind of this generation that has not been modified by Carlyle’s writings; there has hardly been an English book written for the last ten or twelve years that would not have been different if Carlyle had not lived... The extent of his influence may be best seen in the fact that ideas which were startling novelties when he first wrote them are now become common-places. And we think few men will be found to say that this influence on the whole has not been for good...
    (Who didn't start the fire...)
    3/5.
    [Free here]
    [Thinking #3, Theory #2]


  • The Shepherd's Crown (2015) by Terry Pratchett. Don't know if the flatness of this comes from its being Young Adult, or from the smoothened, modern nature of his late Discworld, or from the cortical atrophy. Little of his obliquity and spark to show; it feels like someone else's writing, and no doubt it substantially was. Trades on past power, and what power it was: his witches are pre-modern doctor, social worker, priest, undertaker, and night watch. Came to say goodbye, and I got that after 5 short chapters.
    2/5.
    [Library]
    [Values #3]




(c) Errata Corrige #4 (2013) by Ekaterina Panikanova


In that sense, getting shot in the face with a rubber bullet and tear-gassed is a form of validation, a way for the state to say, “We acknowledge your objections and consider them important enough for us to come out here and shoot you in the face over them.”
nuclearspaceheater


NOV
  • Learning Spark (2011) by Karau. Tool books are difficult to stomach: their contents are so much more ephemeral than other technical books. It often feels like it's not worth it: in 10 years, will it matter? etc. (This is an incredibly high bar to pose, but that's how high my opinion is of the technical pursuits.) O'Reilly soften this blow, occasionally, by enlisting really brilliant authors who bring in the eternal and the broad while pootering around their narrow furrow. (I am incredibly fond of Alan Gates for this, for instance.)

    Spark is the biggest deal by far in my corner of the world and will probably affect your life in minor ways you will never pin down (see O'Neil below).
    3/5.
    [Theory #1, Thinking #1]



  • Curiosity: How Science Became Interested in Everything (2012) by Philip Ball. Actually a history of the early modern origins of science: the long trek via Natural Magic, Alchemy, Neoplatonism, herbalism. The context of discovery is messy. Ball doesn't but in general people make way too much of this fact.

    The received view of scientific history is one-dimensional: you have the superstitious qualitative cretins at one end and the atheistic mathematised moderns at the other. Really it needs at least 5 axes before you get even a basic understanding of the great, great revolution that began to happen around the C16th. I've graphed the intellectual space in the full review.

    3*/5.
    [Library]
    [Data #1, 2, 3]



  • Signifying Rappers (1989) by David Foster Wallace and Mark Costello. The first book on hip-hop? Certainly the first High Academic one. Though, not really a book, as they frequently acknowledge: it's a "sampler". And not expert, as they constantly acknowledge: more than half of it is them pseudo-nervously hedging about being two elite white guys peering into what was then a fairly closed circle. A solid effort too - it knows and guesses and connects more than most critics today, despite the scene being far more ethnically closed, and far less obviously of artistic wealth; despite their often comically mishearing the lyrics; and despite not being able to find anything out about the people behind the music, because no-one returned their calls (until they pretended to be journalists).

    Anyway this has 80pp of recognisably enervated DFW popping off the top of this allocortex, decent fuel for the fire of an admirer, or at any rate the only coal on offer (he was embarrassed by this book, but it is too stylish and enthusiastic to be embarrassing to us):
    Ironies abound, of course, as ironies must when cash and art do lunch. Tearing down the prop-thin symbolic walls, Run-DMC aim to celebrate desegregation, but miss the fact that Aerosmith, those whitest of white rockers, are merely big-budget Led Zeppelin ripoffs, and that Led Zep came straight outta the jet-black Rhythm & Blues of Chicago’s Chess Records. Dancing with Steve Tyler, Run-DMC forgets that Muddy Waters’ sideman Willie Dixon had to sue Led Zeppelin to get proper credit for their use of his blues. “Walk This Way” is an unwanted reunion of 80s black street music with part of its rich heritage, as that heritage has been mined and mongrelized by Show Biz. If this is desegregation, then shopping malls hold treasure...

    It’s a new and carnivorous kind of mimesis that makes weary old ‘self-reference’ actually kind of interesting, because it enlarges Self from the standard rock-subjective–a bundle of hormone-drenched emotions attached to a larynx and pelvis–to a 'big ole head,’ a kind of visual street-corner, a monadic Everybrother, an angry, jaded eye on a centerless pop-culture country full of marginalized subnations that are themselves postmodern, looped, self-referential, self-obsessed, voyeuristic, passive, slack-jawed, debased, and sources of such prodigious signal-and-data bombardment that they seem to move faster than the angry eye itself can see…
    I had been putting off reading this because of the title: I didn't know about Schooly D's track, so I read the verb in a gross academic voice ("in which we give rappers true signification") rather than the adjectival sense they actually meant ("rappers who signify").

    Costello's bits are ok, DJ "MC" to MC "DFW". Wallace is harder than Costello - noting that MCs really are just yuppies, that Chuck D's claims to not be glorifying violence are absurd, that part of the fascination of hard rap is the snuff-spiral of trying to be nastier and nastier than previous hard rappers, which is just the commercial impulse of Alice Cooper minus musicianship. But this is also a winning early bet: that rap is poetry, that it was and would be "the decade's most important and influential pop movement":
    Our opinion, then, from a distance: not only is serious rap poetry, but, in terms of the size of its audience, its potency in the Great U.S. Market, its power to spur and to authorize the artistic endeavor of a discouraged and malschooled young urban culture we’ve been encouraged sadly to write off, it’s quite possibly the most important stuff happening in American poetry today. ‘Real’ (viz. academic) U.S. poetry, a world no less insular than rap, no less strange or stringent about vocal, manner, and the contexts it works off, has today become so inbred (against its professed wishes) inaccessible that it just doesn’t get to share its creative products with more than a couple thousand fanatical, sandal-shod readers..
    Your enjoyment will depend on you giving a crap about the sheer horror of rap's initial context and being able to tolerate intentionally torturous pomo prose and juxtapositions (e.g. I Dream of Jeannie vs race riots). I loved it and twice missed my stop on the tube reading it.
    3*/5.
    [Library]
    [Data #1, Theory #1, Values #1, Thinking #2]





(c) Errata Corrige #11 (2013) by Ekaterina Panikanova

DEC
  • House of God (1971) by Samuel Shem. Updikean satire, more delightful than funny. Its surrealism, puns (Mrs Risenshein, an LOL in NAD [litle old lady in no obvious distress]), sexual glibness earn it a right to sentimentality in the face of human filth and pain:
    We fought. She probably knew we were fighting about Dr. Sanders’s long dying and about the illusion in my father’s letters and about my plethora of absent role models and the blossoming idea that the gomers were not our patients but our adversaries, and most of all we were fighting over the guilt that I felt for having Molly in a dark corner of the ward standing up, this Molly, who, like me, wouldn’t stop and think and feel either, because if she ruminated on what she felt about enemas and emesis basins, she’d lose faith even in her centipede and want to kill herself too. Our fight was not the violent, howling, barking fight that keeps alive vestiges of love, but that tired, distant, silent fight where the fighters are afraid to punch for fear the punch will kill. So this is it, I thought dully, four months into the internship and I’ve become an animal, a mossbrained moose who did not and could not and would not think and talk, and it’s come like an exhausted cancerous animal to my always love, my buddy Berry, and me–yes it’s come to us: Relationship On Rocks...
    Shem's dialogue is pleasurable - the Flann O'Joyce variety of brainy silliness. His two eloquent Irish cops are the best people in the book:
    "Top o' the morning to you, brave Sergeant Finton Gilheeney."
    "Is it the Commissioner?"
    "None other. The young doctor says that with the aid of an operation, with the usefulness of the scalpel being demonstrated, you will survive."
    "-Dr Basch, I believe that I now have no need of the last rites. If so, could the priest depart? He scares me in the memory of how close to heaven or that hot other place I came."
    "-And is there a message for the little woman, the wife?" the Commissioner asked as the priest left...

    "Ah well, all the best boyo, and I'm on my way to your wife and will soothe her with my boyish good looks and TV-cop mien. Good-bye, and for the young scholar here who saved your fine red life, SHALOM and God bless."

    Savage, all of it, savage.
    Like any psychologically ambitious work of the mid-C20th, it has a lot of Freud in it, much of it going unchallenged. The book is also about the distress and pain of an extremely lucky and insulated and remunerated man surrounded by women who do massive amounts for him, but you mostly forget this, it is that good.

    I imagine there are still pockets of people out there who still believe in the 1950s George Clooney heroism and omnicompetence of doctors. So Shem, hot-shot prof at BMS, and his book have work to do.
    4/5.
    [Theory #2, Values #2]


  • Seveneves (2014) by Neal Stephenson. Amazing hard worldbuilding from a lunatic seed: 'what would happen if the moon just blew up?' You will stomach pages of physical exposition before most scenes, but none of it is superfluous. First two-thirds are psychologically convincing: you will ball your fists at the politics. Couple of railroaded plot points - e.g. it is taken for granted that a psychopathic war criminal has every right to an equal share of the genetic future. And the last third's extrapolation of 5000 years of cultural creep is less formally ambitious than e.g. Cloud Atlas.

    First two-thirds 4*/5, last third 3/5.
    [Library]
    [Theory #1, Theory #2, Theory #4, Values #2]


  • Weapons of Math Destruction (2016) by Cathy O'Neil. Original, important, expert, impassioned in the right places. I have some gripes of course; full review here.
    4*/5.
    [Data #2, Theory #1, Theory #3, Values #1]


  • Born to Run (2016) by Bruce Springsteen. Fans only. Though you probably will be one if you've given him the time: he is unusual among rock auteurs, populist and wholesome to the point of naivete:
    I was... a circumstantial bohemian - I didn't do any drugs or drink... I was barely holding on to myself as it was. I couldn't imagine introducing unknown agents into my system. I needed control and those ever-elusive boundaries... Music was going to get me as high as I needed to go... the counterculture stood by definition in opposition to the conservative blue-collar experience I'd had.
    Prose is clumsy enough to be actually his work, and is eloquent by rockstar standards:
    When it rains, the moisture in the humid air blankets our town with the smell of damp coffee grounds wafting in from the Nescafé factory at the town’s eastern edge. I don’t like coffee but I like that smell. It’s comforting; it unites the town in a common sensory experience; it’s good industry, like the roaring rug mill that fills our ears, brings work and signals our town’s vitality. There is a place here—you can hear it, smell it—where people make lives, suffer pain, enjoy small pleasures, play baseball, die, make love, have kids, drink themselves drunk on spring nights and do their best to hold off the demons that seek to destroy us, our homes, our families, our town...
    He's had thirty years of psychotherapy, the poor sod. He is intellectual enough to take his feelings and their theories seriously - but not intellectual enough to be sceptical about their interminable unscientific faffing.
    2/5.
    [Values #3]


  • Keeping On Keeping On (2016) by Alan Bennett. Diaries in the lee of becoming actually famous. I love him dearly and bolted all 700pp in a couple of days. General sense of him reaping decades of quiet acclaim: he bumps into well-wishers and heavy-hitters (Stoppard, ) every week or so. One of the reasons I love him is that I had a very similar adolescence to his. And he remains a reserved sort, kind-but-grumpy:
    Being in love unhappily singled you out, I thought, it drafted you into an aristocracy. It was more than just a badge of being gay but rather an ordeal you were called upon to undergo if only to transcend it and reach a sublimity denied to other mortals.

    In the evening to the New York Public Library where I am to be made a Library Lion... There are half a dozen of us being lionised and we are lined up and photographed and made much of before going upstairs to a magnificent supper, getting home thoroughly knackered around 11. How people lead a social life is beyond me.

    I clung far too long to the notion that shyness was a virtue and not, as I came too late to see, a bore.
    He still feels outside of things, for all his reminiscences of dinner with Harold Wilson or Liz Taylor perching on his knee. On winning a Tony for Best Play aged 72: donoho
    I am thrust blinking on to a stage facing a battery of lights while questions come out of the darkness, the best of which is: ‘Do you think this award will kick-start your career?'
    Talks so much about 50s Yorkshire. (People in general seem to think about their childhoods more than I do. (or just writers?)) I suppose he is taken to be a twee writer for this nostalgia, along with his cuddly speaking voice. But he simply isn't twee - he is the author of several of the finest nihilist soliloquoys in English literature. You may know the ignorance of people by their use of this stereotype.

    He is touchingly agitated by British politics, in the exact way I used to be. His protests are unprogrammatic, based simply on the meanness or indignity or cowardice of the policy at hand, whether it's a Labour or Tory hand;
    I wanted a Labour government so that I could stop thinking about politics, knowing that the nation’s affairs were in the hands of a party which, even if it was often foolish, was at least well-intentioned. Now we have another decade of the self-interested and the self-seeking, ready to sell off what’s left of our liberal institutions and loot the rest to their own advantage. It’s not a government of the nation but a government of half the nation, a true legacy of Mrs Thatcher...

    I’ve always thought that this was a pretty fair description of that blend of backward-looking radicalism and conservative socialism which does duty for my political views. I am an old modernian... [Over the past 30 years] one has only had to stand still to become a radical.
    With the fading of the old loud left, and the abject failure of the sneering theoretical sort, unpretentious justice of this sort might motivate people, even/especially opportunist Brexiters. So to the defence of public libraries, the unprecedented conviction of policemen who murder, the provision of good to all.
    4*/5.
    [Library]
    [Data #1, Values #3, Thinking #3]


  • On the Move (2015) by Oliver Sacks. Rushed: just a string of events and bad prose extracts from his adolescence. Also two long chapters exaggerating the achievements of two scientific titans vs consciousness studies (Crick and Edelman). Hadn't known his love life was so fraught - he looks like such a bull (and indeed Bennett remembers Sacks at Oxford as a brash alpha). Weightlifting chat is endearing in an intellectual. Read his real books.
    3/5.
    [Values #3, Theory #1]




Acause incomer will ayeways be a clarty wird
acause this tongue A gabber wi will nivver be the real Mackay, A sing
Acause fer aw that we’re aw Jock Tamson’s etcetera, are we tho? Eh? Are we.
Acause o muntains, castles, tenements n backlans,
acause o whisky exports, acause o airports,
acause o islans, A sing.
acause of pubs whit arena daein sae weel oot o the smokin ban, A sing.
a cause hit's grand tae sit wi a lexicon n a deeskit mynd, A sing.
acause o the pish in the stair, A sing.
acause o ye,

A sing o a Scotland whit wadna ken working class authenticity gin hit cam reelin aff an ile rig douned six pints n glasst hit in the cunt.
whit hit wadna
by the way.
A sing o a google Scotland
o laptop Scotland
o a Scotland saw dowf on bit-torrentit HBO
drama series n DLC packs fer paistapocalyptic RPGs that hit wadna ken
hits gowk fae its gadjie,
fae whas lips n fingers amazebawz cams
mair freely as bangin.
...
A sing o a Scotland bidin in real dreid o wan day findin oot
juist hou parochial aw hits cultural references mey be,
n cin only cope wi the intertextuality o the Scots
Renaissance wi whappin annotatit editions,
n weens hits the same wi awbdy else.
I sing o a Scotland whit’ll chant hits hairt oot dounstairs o the Royal Oak, whit’ll pouk hits timmer clarsach hairtstrangs, whit like glamour will sing hits hairt intae existence, whit haps sang aroon hits bluidy nieve hairt,
whit sings.
- Harry Giles



27/12/2016

Checklist for toxic algorithms


Based on comments in O'Neil's Weapons of Math Destruction. Full review here.

Opacity
  • Is the subject aware they are being modelled?
  • Is the subject aware of the model's outputs?
  • Is the subject aware of the model's predictors and weights?
  • Is the data the model uses open?
  • Is it dynamic - does it update on its failed predictions?

Scale
  • Does the model make decisions about many thousands of people?
  • Is the model famous enough to change incentives in its domain?
  • Does the model cause vicious feedback loops?
  • Does the model assign high-variance population estimates to individuals?

Damage
  • Does the model work against the subject's interests?
    • If yes, does the model do so in the social interest?
  • Is the model fully automated, i.e. does it make decisions as well as predictions?
  • Does the model take into account things it shouldn't?
  • Do its false positives do harm? Do its true positives?
  • Is the harm of false positives symmetric with the good of true positives?



Note that "Inaccuracy" is not a criterion for O'Neil. This is maybe the core shortcoming of the book: it doesn't wrestle much with the hard tradeoff involved in when modelling unfair situations, e.g. living in a bad neighbourhood which increases your risks and insurance costs through no fault of your own. She comes down straightforwardly on the direct "make the model pretend it isn't there" diktat. My preferred measure would be to not prevent models from being rational, but instead make transfers to the victims of empirically unfair situation. (This looks pointlessly indirect, but price theory, and the harms of messing with them, is one of the few replicated economic.) My measure has the advantage of not requiring a massive interpretative creep of regulation: you just see what the models do as black boxes and then levy justice taxes after.


11/12/2016

Curiosity: a new fragrance by Philip Ball.

...
—why is the sea salty?
—have animals souls, or intelligence?
—has opinion its foundation in the animate body?
—why do human beings not have horns?
—how is it that sound in its passage makes its way through any obstacle whatever?
—how is it that joy can be the cause of tears?
—why are the fingers of unequal length?
—why, if you have intercourse with a woman after she has lain with a leper, will you catch the disease while she will escape?
—what reason is there for the universality of death?
—why do we need food so frequently, or at all?
—why are the living afraid of the bodies of the dead?
—how is the globe supported in the middle of the air?
—why does the inflow of the rivers not increase the bulk of the ocean?
—why, if a vessel be full and its lower part open, does water not issue from it unless the upper lid be first removed?
—when one atom is moved, are all moved? (since whatever is in a state of motion moves something else, thus setting up infinite motion.)
—why do winds travel along the earth's surface and not in an upward direction?
—why does a sort of perpetual shadow brood over the moon?
—granted that the stars are alive, on what food do they live?
—ought we regard the cosmos as an inanimate body, a living thing, or a god?

— Adelard of Bath (c.1120)


Curiosity: How Science Became Interested in Everything (2012) by Philip Ball.

Another history of the origins of science: our long trek to GWAS, livermorium, and CERN via astrology, natural magic, alchemy, Neoplatonism, herbalism, occultism, and philosophy. So, superficially, the book is just about an especially fruity context of discovery. But this period holds two of the most important lessons in history: 1) science grew out of work by people who diverge wildly from the modern idea and practice of science, whose variously false frameworks led to the Royal Society and e.g. the Newtonian triumph. (And from there to contemporary, professional, university science.) So wrong people can still make progress if their errors are uncorrelated with the prevailing errors. And, 2) a small number of the most powerful people in Britain - the Lord Chancellor, the king's physicians, the chaplain of the young Elector Palatine and bishop of Chester, London's great architect, Privy Councillors * - successfully pushed a massive philosophical change, and thereby contributed to most of our greatest achievements: smallpox eradication, Sputnik and Voyager, the Green Revolution, and the unmanageably broad boons of computing are partly theirs.

The received view of all this is one-dimensional: you have superstitious, pompous cretins at one end and rational, experimental moderns at the other.

But really you need five axes before you get a basic understanding of the great, great revolution that began in the C16th - before you can see how science differs from every other community:

  • Supernaturalism vs Naturalism. Did they explain things solely in terms of natural causes? (Absentee Gods only.)

  • Apriori vs Aposteriori. Did they view actual observation as decisive and indispensable? **

  • Qualitative vs Quantitative. Did they make measurements? Did they model the data? Did they use standard units?

  • Holism vs Reductionism. Did they analyse things into their constituent features? Did they explain phenomena in terms of ?

  • Infallibilism vs Fallibilism. Did they allow for the possibility of error? Did they view uncertain knowledge as still worthwhile? ***

  • ****
^
So I'm modelling science as naturalist, fallibilist, quantitative empiricism with pretensions to openness. I've categorised the early scientists mentioned in Curiosity according to this: you can see the data with additional justifications here. (Ball doesn't state this model, but it floats around in his debunkings and "well actually"s.)




All of the pieces of science are very ancient - we had mathematics and data collection well before the Ten Commandments, naturalism before Buddha and Confucius, reductionism before the Peloponnesian War at least one controlled trial centuries before Christ, fallibilism likewise. Everything was ready BCE; we can see indirect evidence of this in the astonishing works of Ancient Greek engineers, mostly unmatched for 1000 years until y'know.

So the question is not "was Bacon the most original blah blah?": he wasn't, particularly when you remember Alhazen's Baconian method, developed in the C11th. But we need an explanation for how we messed it up so badly. The received view, which is all I have at the moment, is that the fall of Rome, Christian anti-intellectualism and, later, the enshrining of Aristotelian mistakes was enough to destroy and suppress the ideas. I want deeper explanations though. (For instance, what did we do to the economy?)


A fun regression on this data would be to see how my scienciness measure correlates with the importance of the person's work. It would not be that highly proportional, in this time period.


***************************************************************

Back to the book eh!

Book structure is lots of little chapters on fairly disjointed topics: early modern ideas of space travel, universal language, pumps, etc. Chapter on "cabinets of curiosity" is great though: suddenly their dull zany blare makes sense and I want to build one:
this was more than a case of 'look what I've got'. The power with which Wunderkammern were imbued was... in that they created their own complete microcosm: a representation of the world in miniature... By possessing this microcosm the collector-prince was not just symbolising but also in a sense exercising his mastery of the world. The cabinet acted as a kind of mental laboratory within which the relationships between things could be contemplated via a process that shared elements of both experimentation and Gnostic revelation.
Ball doesn't like us calling the Scientific Revolution a revolution, and I agree: the revolution didn't consist in the theories of Bacon or Newton: it consists in the diffusion of the worldview into all subjects and all inquiry. It transformed society and gave us marvels, but it hasn't finished happening. The general will, or default state, is still strongly unscientific. (The largest and most grievous holdout, larger even than the enduring hold of fideist religion, is our tribal politics and our largely nonempirical government policy.)

Ball expends a lot of time on a history of wonder vs curiosity vs dispassionate robot inquiry. People hated all of these things for various reasons, up until the Renaissance when curiosity became acceptable on what are now classic economic grounds, or in line with the Italian cult of the virtuoso - someone who's so bloody brilliant that you have to just let him get on with it.

I always like Ball's drawling prose and catty editorialising. (For instance, Margaret Cavendish - the darling of arts academics who latch on to the only woman in sight in this period - gets a round dissing by Ball, as an anti-experiment idiot, a vitalist, and a misogynist.) Stimulating as always.

4/5. [Library]



***********************************************************************************


* Bacon has some claim to being the most influential philosopher ever, in terms of counterfactual effect on history. (Rather than number of bloody citations!) No-one with his social standing was resisting the Aristotelian consensus in 1620; his prototype scientific method is a century ahead of its time. (Yes, ibn al-Haytham's was 7 centuries ahead of its time, but to limited avail.)


** This one is hard to refer to, because we now find it incredibly easy to understand why "go and look" works as a general route to knowledge; Medieval thought rejected this on the basis of things like the problem of induction.

The cliched way to refer to the split between those who want to start with the apriori and those who want to start with data is "Rationalism" vs "Empiricism". But these words confuse people: the two of them are also used in a C17th debate about psychology, to do with the nature of mental content.

More: it can't be a dichotomy, since many of the greatest rationalists (Descartes, Leibniz) were experimentalists too, doing what we now call empirical work. Three meanings of rationalism, and three words for them:
  • 'Rationalism1': Belief in innate ideas. Call it 'Continental Rationalism'. Descartes and Leibniz but not Dawkins and Shermer.
  • 'Rationalism2': Belief in the supremacy of apriori knowledge over empirical knowledge. Call it 'apriorism'. Aristotle was apriorist, as was Descartes.
  • 'Rationalism3': Belief that everything should be subject to reason and evidence. Includes Descartes and Leibniz and Dawkins and Shermer. Contemporary rationalists are highly if not radically empiricist.

I use Alberto Vanzo's criteria for deciding if someone was enough of an experimentalist:
let us consider four typical features of early modern experimental philosophers:
  • self-descriptions: experimental philosophers typically called themselves such. At the very least, they professed their sympathy towards experimental philosophy.
  • friends and foes: experimental philosophers saw themselves as part of a tradition whose “patriarch” was Bacon and whose sworn enemy was Cartesian natural philosophy.
  • method: experimental philosophers put forward a two-stage model of natural philosophical inquiry: first, collect data by means of experiments and observations; second, build theories on the basis of them. In general, experimental philosophers emphasized the a posteriori origins of our knowledge of nature and they were wary of a priori reasonings.
  • rhetoric: in the jargon of experimental philosophers, the terms “experiments” and “observations” are good, “hypotheses” and “speculations” are bad. They were often described as fictions, romances, or castles in the air.
This is unusually inclusive: the famous Rationalist Leibniz counts as experimental under this rubric. But a stronger definition of aposteriorist - like "refuses to use purely analytic reasoning", or even "spent most of their time running experiments and analysing data" would exclude many contemporary scientists. Sticking with Vanzo for now.


*** Hard to imagine a fallibilist apriorist: perhaps Lakatos. (Some say Leibniz was, in practice.) I actually have met a methodist infallibilist apriorist, but I won't meet another.


**** I had included "openness" in the model -
  • Obscurantism vs Openness. Did they write in the vernacular? Did they publish for a wide readership? Did they spurn Noble Lies? Did they encourage replications with and data sharing? Did they build scholarly networks?
- but I admit this is just wishful/normative thinking: modern academic science fails at this. Whether with its low-status replications, unreadable prose, paywalls on most research (tax-funded or no), pathetically low levels of data sharing, or the prevalence of noble lies... But it's definitely a core aspiration now: the greedy impulse behind hermeticism is blatantly unscientific, if not actually shunned by actual scientists. First, lip service...

Things can be science without being published, obviously: consider the invention of public key cryptography by a GCHQ wonk, classified for 25 years - or even the secret infrastructure and algorithmics of high-frequency trading.



^ Obviously these five factors aren't the end of the matter either. But I reckon it catches a decent amount of the variance in the term "scientist". Others e.g.

  • Particularism vs Consilience. Did they believe that the scientific method could explain every phenomenon?
  • Realism vs Instrumentalism. Most scientists are realists about best current theories
  • Theism vs Nontheism

I had included non-theism in the core of modern science - and so it is, in the form of strong naturalism. Scientists, on the other hand, differ from this, globally. This is partially because humans are so compartmentalised and can hold severe contradictions indefinitely. But, clearly, atheism is not an essential part of the modern method. But causal closure and (at most) a private faith are.

09/12/2016

notable wordwordword


  • dragon-king (n.): An extreme event among extreme events: roughly, an outlier of a Pareto distribution, even. An elaboration on Taleb's black swan metaphor for unforeseeable extreme events. Not sure if it adds much, since the black swan is distribution-independent and Taleb doesn't fixate on power laws iirc.

  • chef's arse (n.): Painful chafing of the buttocks against each other; attends exercise in hot environments.

  • groufie ( n.): group selfie, obvs. No less contemptible for the awkward swerve around "groupie".

  • detaliate (mangled v.): To explain. Seen in this Quora answer by a non-native English speaker (possibly Romanian). I want to appropriate it: to detaliate is to respond to casual comments with a fisking.

  • consing: (n.): To save on memory allocation by comparing new values to existing allocations and just storing a hash to the existing one if it's a hit. From Lisp's cons cells, a basic key-value data structure.

  • sadcore: ( n.): Slow indie. Journo term: avoided by anyone musical for obvious reasons.

  • key-signing party (n.): A meetup in person for secure exchange of cryptographic keys. Not to be confused with the other key party, but I hope someone has.

  • shitlib (Internet n.): Shit liberal - but clearly not as in "someone who is failing to live up to liberal ideals". Seems to be used by rightists for "advocate of political correctness", though of course leftists attack liberals for defending freedom of speech ("freeze peach").

  • ¯\_(ツ)_/¯


  • statecuck (Internet n.): A supporter of state intervention in something. I was never able to take "cuck" seriously as an insult, but I listen even less since learning it is pronounced "cook".

  • channish ( n. or adj.): the language of 4chan (8chan, etc.); or 2) the property of being fit only for 4chan. Characterised by exaggeration, cruelty, and crude compound neologisms. "Socdem shitlib" and "statecuck" are channish, obviously.

  • Wadsworth constant (n.): The first 30% of a Youtube video. The part of a video that can be skipped because it will contain "no worthwhile or interesting information".

  • Kegan levels ( n.): the ranked parts of a particular theory of consciousness. I usually disdain these things, since the data has been so poor and the theorising simultaneously fragile and unfalsifiable. But I find this one really interesting, not least for its falsifiability - e.g. "logical reasoning is a necessary condition of having stable feelings", "most people never reach this stage, and never before age 40".

  • shirt-tail relatives ( n.): Can't beat this Quora answer: "people related to me enough that someone in my family circle cares deeply about their lives, but not closely enough related to me that I have any idea what the relationship is, in a biological or legal sense."

  • annalist (pej. n.): An insult between historians: someone who merely records events uncritically. Comes from the official scribes of ancient Rome, who were indeed uncritical to the point of deceit.



21/11/2016

Notes from Effective Altruism "Global" "x" in Oxford, in 2016


(This is about this thing. The following would work better as a bunch of tweets but seriously screw that: )


###########################################################################################################


Single lines which do much of the work of a whole talk:

"Effective altruism is to the pursuit of the good as science is to the pursuit of the truth." (Toby Ord)

"If the richest gave just the interest on their wealth for a year they could double the income of the poorest billion." (Will MacAskill)

"If you use a computer the size of the sun to beat a human at chess, either you are confused about programming or chess." (Nate Soares)

"Evolution optimised very, very hard for one goal - genetic fitness - and produced an AGI with a very different goal: roughly, fun." (Nate Soares)

"The goodness of outcomes cannot depend on other possible outcomes. You're thinking of optimality." (Derek Parfit)



###########################################################################################################


Owen Cotton-Barratt formally restated the key EA idea: that importance has a highly heavy-tailed distribution. This is a generalisation from the GiveWell/OpenPhil research programme, which dismisses (ahem, "fails to recommend") almost everyone because a handful of organisations are thousands of times more efficient at harvesting importance (in the form of unmalarial children or untortured pigs or an unended world).

Then, Sandberg's big talk on power laws generalised on Cotton-Barratt's, by claiming to find the mechanism which generates that importance distribution (roughly: "many morally important things in the world, from disease to natural disasters to info breaches to democides all fall under a single power-law-outputting distribution").

Cotton-Barratt then formalised the "Impact-Tractability-Neglectedness" model, as a precursor to a full quantitative model of cause prioritisation.



Then, Stefan Schubert's talk on the younger-sibling fallacy attempted to extend said ITN model with a fourth key factor: awareness of likely herding behaviour and market distortions (or "diachronic reflexivity").

There will come a time - probably now tbf - when the ITN model will have to split in two: into one rigorous model with nonlinearities and market dynamism, and a heuristic version. (The latter won't need to foreground dynamical concerns unless you are 1) incredibly influential or 2) incredibly influenceable in the same direction as everyone else. Contrarianism ftw.)


###########################################################################################################


Catherine Rhodes' biorisk talk made me update in the worst direction: I came away convinced that biorisk is both extremely neglected and extremely intractable to anyone outside the international bureaucracy / national security / life sciences clique. Also that "we have no surge capacity in healthcare. The NHS runs at "98%" of max on an ordinary day."

(This harsh blow was mollified a bit by news of Microsoft's mosquito-hunting drones - used for cheap and large-sample disease monitoring, that is, not personalised justice.)


###########################################################################################################




Anders Sandberg contributed to six events, sprinkling the whole thing with his hyper-literate, uncliched themes. People persisted in asking him things on the order of "whether GTA characters are morally relevant yet". But even these he handled with his rigorous levity.

My favourite was his take on the possible expanded value space of later humans: "chimps like bananas and sex. Humans like bananas, and sex, and philosophy and competitive sport. There is a part of value space completely invisible to the chimp. So it is likely that there is this other thing, which is like whoooaa to the posthuman, but which we do not see the value in."


###########################################################################################################


Books usually say that "modern aid" started in '49, when Truman announced a secular international development programme. Really liked Alena Stern's rebuke to this, pointing out that the field didn't even try to be scientific until the mid-90s, and did a correspondly low amount of good, health aside. It didn't deserve the word, and mostly still doesn't.


###########################################################################################################


Nate Soares is an excellent public communicator: he broadcasts seriousness without pretension, strong weird claims without arrogance. A catch.


###########################################################################################################


What is the comparative advantage of us 2016 people, relative to future do-gooders?

  • Anything happening soon. (AI risk)
  • Anything with a positive multiplier. (schistosomiasis, malaria, cause-building)
  • Anything that is hurting now. (meat industry)


###########################################################################################################


Dinner with Wiblin. My partner noted that I looked a bit flushed. I mean, I was eating jalfrezi.


###########################################################################################################


Most every session I attended had the same desultory question asked: "how might this affect inequality?" (AI, human augmentation, ) The answer's always the same: if it can be automated and mass-produced with the usual industrial speed, it won't. If it can't, it will.

It was good to ask (and ask, and ask) this for an ulterior reason though, see the following:


###########################################################################################################


Molly Crockett's research - how a majority of people* might relatively dislike utilitarians - was great and sad. Concrete proposals though: people distrust people who don't appear morally conflicted, who use physical harm for greater good, or more generally who use people as a means. So express confusion and regret, support autonomy whenever the harms aren't too massive to ignore, and put extra effort into maintaining relationships.

These are pretty superficial. Which is good news: we can still do the right thing (and profess the right thing), we just have to present it better.

(That said, the observed effects on trust weren't that large: about 20%, stable across various measures of trust.)



* She calls them deontologists, but that's a slander on Kantians: really, most people are just sentimentalists, in the popular and the technical sense.



###########################################################################################################


Not sure I've ever experienced this high a level of background understanding in a large group. Deep context - years of realisations - mutually taken for granted; and so many shortcuts and quicksteps to the frontier of common knowledge. In none of these rooms was I remotely the smartest person. An incredible feeling: you want to start lifting much heavier things as soon as possible.


###########################################################################################################



Very big difference between Parfit's talk and basically all the others. This led to a sadly fruitless Q&A, people talking past each other by bad choice of examples. Still riveting: emphatic and authoritative though hunched over with age. A wonderful performance with an air of the Last of His Kind.

Parfit handled 'the nonidentity problem' (how can we explain the wrongness of situations involving merely potential people? Why is it bad for a species to cease procreating?) and 'the triviality problem' (how exactly do tiny harms committed by a huge aggregate of people combine to form wrongness? Why is it wrong to discount one's own carbon emissions when considering the misery of future lives?).

He proceeded in the (lC20th) classic mode: state clean principles that summarise an opposing view, and then find devastating counterexamples to them. All well and good as far as it goes. But the new principles he sets upon the rubble - unpublished so far - are sure to have their own counterexamples in production by the grad mill.


The audience struggled through the fairly short deductive chains, possibly just out of unfamiliarity with philosophy's unlikely apodicticity. They couldn't parse it fast enough to answer a yes/no poll at the end. ("Are you convinced of the non-difference view?")

The Q&A questions all had a good core, but none hit home for various reasons:

  • "Does your theory imply that it is acceptable to torture one person to prevent a billion people getting a speck in their eye?" Parfit didn't bite, simply noting, correctly, that 1) Dostoevsky said this in a more manipulative way, and 2) it is irrelevant to the Triviality Problem as he stated it. (This rebuffing did not appear to be a clever PR decision, though it was, since he is indeed a totalarian.)

  • "What implications does this have for software design?" Initial response was just a frowning stare. (Sandberg meant: lost time is clearly a harm; thus the designers of mass-market products are responsible for thousands of years of life when they fail to optimise away even 1 second delays.)

  • "I'd rather give one person a year of life than a million people one second. Isn't continuity important in experiencing value?" This person's point was that Parfit was assuming the linearity of marginal life without justification, but this good point got lost in the forum. Parfit replied simply - as if the questioner was making a simple mistake: "These things add up". I disagree with the questioner about any such extreme nonlinearity - they may be allowing the narrative salience of a single life to distract them from the sheer scale of the number of recipients in the other case - but it's certainly worth asking.


We owe Parfit a lot. His emphasis on total impartiality, the counterintuitive additivity of the good, and most of all his attempted cleaving of old, fossilised disagreements to get to the co-operative core of diverse viewpoints: all of these shine throughout EA. I don't know if that's coincidental rather than formative debt.

(Other bits are not core to EA but are still indispensable for anyone trying to be a consistent, non-repugnant consequentialist: e.g. thinking in terms of degrees of personhood, and what he calls "lexical superiority" for some reason (it is two-level consequentialism).)

The discourse has definitely diverged from non-probabilistic apriorism, also known as the Great Conversation. Sandberg is of the new kind of philosopher: a scientific mind, procuring probabilities, but also unable to restrain creativity/speculation because of the heavy, heavy tails here and just around the corner.



17/11/2016

data jobs, tautologies, bullshit, $$$


(c) Tom Gauld (2014)
When physicists do mathematics, they don’t say they’re doing “number science”. They’re doing math. If you’re analyzing data, you’re doing statistics. You can call it data science or informatics or analytics or whatever, but it’s still statistics... You may not like what some statisticians do. You may feel they don’t share your values. They may embarrass you. But that shouldn’t lead us to abandon the term “statistics”.

Karl Broman


what makes data science special and distinct from statistics is that this data product gets incorporated back into the real world, and users interact with that product, and that generates more data: a feedback loop. This is very different from predicting the weather...

– Cathy O'Neil / Rachel Schutt


"Data science" is the latest name for an old pursuit: the attempt to make computers give us new knowledge. * In computing's short history, there have already been about 10 words for this activity (and god knows how many derived job titles). So: here's an anti-bullshit exercise, a genealogy of some very expensive buzzwords.

The following are ordered by the year the hype peaked (as estimated by maximum mentions in books). You can play with proper data here.




  • "Expert systems"
    The original, GOFAI craft. Painstaking, manually-built rule stacks. 69% accuracy in certain medical tasks, which beat out human experts.


  • "Business intelligence"
    The most transparently hokum. I include the rigorous but dead-ended world of MDX and OLAP in here, perhaps unfairly: they're certainly still in use by some organisations who you'd expect to know better.


  • "Data mining". Originally a pejorative, among actual statisticians, meaning "looking for fake patterns to proclaim". Now reclaimed in industry and academia.** Compared to ML, data mining has a lot of corporate dilution, proprietary gremlins and C20th crannies in, from what I can tell. (Basically the same as "Knowledge discovery"?)


  • "Predictive analytics". See machine learning but subtract most.


  • "Big data". Somewhat meaningful as a concept, extremely tangible as an engineering challenge, and tied to genuinely new results. But still highly repugnant. Has captured much of the present job market, but the hype train has headed off well and truly.


  • "Machine learning". Applied statistics, but recast by computer scientists into algorithms. Goal: Getting systems that work fast rather than inferring the calibrated convergent truth. Along with stats, ML is the heart of the actual single phenomenon underlying all this money and hype.


  • "Data science". Recent high-profile successes in the AI/ML/DS space are largely due to the data explosion - not to new approaches or smarter protagonists. So this is at least half job title inflation. Still, it is handy to have a job title with enough elbow-room to be statistician and developer and machine teacher at once.


You might have hoped that nominally scientific minds would shun the proliferation of tautologous or meaningless terms. But stronger pressures prevail - chiefly the need of job security, via bamboozling clients or upper management or tech conference attendees.



##################################################################################################

* As always, we settle for optimal guesses instead of 'knowledge'.

If I'd said "the attempt to get knowledge from data" then of course I would just be describing statistics.^ This near miss doesn't bother me - despite the fact that statisticians computerised before any other nonengineering profession or field - and despite their building much of the theory and even implementations described in this piece (besides expert systems and GOFAI). Their gigantic century of work is a superset of what I'm talking about.
^ Obviously my initial definition is pretty close to "narrow artificial intelligence" too: at the limit, AI is "building a system for automatically getting knowledge from arbitrary input". Many of the successes described above also belong to them (particularly expert systems and GOFAI). "Data jobs", as I blandly put it in the header, are "jobs dealing with the fact that we don't quite have AI". There are a lot of terrible data jobs, and I'm not talking about them either. The full specification, if you bloody insist, is: "cultures, largely applied or industrial ones, which use cool data processing methods which are not really A.I. in the wide or strong sense, but which aren't standard 70s drone analyst work either. Nor have they anything to do with the very similar work of information physicists or electronic engineers or anything."

(But then all work in applied maths and stats shares a lot, since it's all based on the same world using the same concepts and logics. Only the goals and technologies really vary.)

I'm speaking as generally as I am - that is, almost speaking nonsense - so I can cut through the mire of terms, the effluent of the academic-industrial complex. In intellectual terms, it is pretty easy to refer to all the things I am trying to refer to: they are 'the formal sciences'. But I'm trying to tease out the practitioners, and the way-downstream economics.


** I had been dismissing "data mining" as just a 90s business way of saying "machine learning", but the distinction is actually fairly well-defined:
Data mining: Direct algorithm design for already well-defined goals - where you know what features to use. (e.g. "What kind of language do CVs use?")

Machine learning: Indirect algorithm design, via automated feature engineering, for a ill-defined goal. (e.g. "How do we distinguish a picture of a cat from a picture of a far away lynx ?")