Here’s my Saturday weekly reads. Links to what I enjoyed reading, with commentary.
1. AI GPUs crush Moore’s law. Human intuition appears to be hardwired to linear growth. Steve is a teenager, and if he grows an inch/year he’ll soon wind up as tall as his dad. But what if Steve grew exponentially? He’d match his dad at 6 feet tall in one year, be 12 feet tall next year, a mile high in 10 years, 1000 miles high in 20 years.
The point is an intuition resisting post by Dario Amodei and Danny Hernandez on exponential improvement in compute capability for training AIs. The doubling time is 3.5 months! Again, 3.5 months!! Compare that to 18 months for Moore’s Law (lame). Quote: “Since 2012, this metric has grown by more than 300,000x (an 18-month doubling period would yield only a 12x increase).” This is driven by GPU improvements and the increasing scale of AI investment. Also “We see multiple reasons to believe that the trend in the graph could continue.” Our minds are not built to grasp the implications of exponential growth. If nothing else, I do think it helps to internalize that the rate of improvement is far larger than Moore’s law. One consequence is we’ll see large progress in AI even if we assume (unlikely) no improved AI algorithms. If you click through the link, make sure to flip back and forth between linear scale and log scale on the graph, which brings the point home. link
2. Google Duplex skepticism. Here’s a story to pair with the above. The recorded demo of Google Duplex shown last week booking a hair appointment and restaurant reservation was very cool. But via John Gruber’s site it’s clear that demo was partly staged/edited. Leaving me more skeptical on whether this is demoware or a product ready for the real world. Caveats are: i) person answering the phone doesn’t give the name of the place, ii) there’s no ambient noise, iii) they didn’t ask for phone number/name of the person calling to book the appointment. Axios story. Gruber skepticism. Gruber dinging Mashable for stealing the story and getting it wrong. Also, finding the restaurant used in the demo via a twitter ask (which is interesting by itself).
3. Europa plumes. Besides Earth, the only known liquid water in our solar system is on Jupiter’s moon Europa and Saturn’s moon Enceladus. These moons are ice covered, with tidal forces keeping the watery oceans under the ice crust at liquid temperature. Enceladus is known to have plumes of water ice spurting out into space, when tidal forces occasionally crack the surface. But Europa is much bigger than Enceladus, with a miles deep ice crust that might not crack. So perhaps no plumes. But a new study based on 20 year old (newly analyzed) Galileo space probe data found a plume after all. In particular by matching a magnetic field change and increase in electron density with Hubble imagery. Excellent! NASA is planning to launch the Europa Clipper mission in 2022. And I was sort of bummed they didn’t choose Enceladus, since plumes seem to be the best hope for finding microbial life in our solar system. Now it’s all good. link
4. Ice core lead from Rome. New paper analyzes ice cores from Greenland, tracking lead pollution from Rome. The new study tracks 25,000 measurements, compared to 18 from a previous one from the 1990s. So it’s way better. The best part is the chart below, mapping a timeline of Roman history onto lead pollution. link
5. Homo naledi had a small but complex brain. New paper. Quote: “Despite its small brain size, H. naledi shared some aspects of human brain organization, suggesting that innovations in brain structure were ancestral within the genus Homo.” Human brains run around 1450 cc, Chimp’s 350, Gorilla’s 530, H. naledi was 510. So very small for the Homo genus.
What this brings to mind is Kevin Laland’s book Darwin’s Unfinished Symphony, which argues human evolution got kickstarted by getting into a virtuous cycle of cultural transmission. So brains evolved not to think, but to copy. Humans know not why we succeed. We merely copy those we believe are successful. Who in turn also don’t know why they do what they do. In this model, we first evolved a brain that was organized to culturally learn, and only later did that put selective pressure on making brains larger. So the sequence is 1. evolutionary selection for cultural transmission -> 2. modified brain (still small) to get better at cultural transmission -> 3. selection for bigger brains for better cultural transmission -> 4. language evolves to transmit culture better -> 5. culturally driven technology improvement.
Along those lines, I ran across this recent paper. From the abstract: “we focus on the growing body of work that treats language structure as emerging from the process of cultural transmission. We argue that a full recognition of the importance of cultural transmission fundamentally changes the kind of questions we should be asking regarding the biological basis of language structure.”
Minor aside. The careful reader will note how many quirks of human behavior are explained once you realize cognition evolved not to be rational, but to culturally imitate successful members of our preferred tribe. Thus my blog, and indeed much of contemporary politics.
It’s Saturday again, so here’s links/commentary on what I read and learned from this week.
1. Begun, the spam war has. The big tech news this week was the announcement of Google Duplex. Duplex launched with a demo of it calling a hair salon on the phone and talking to a human to make an appointment. If you haven’t seen it yet, pause and look, it’s worth watching. The computer sounds very human.
What this brought to mind is an odd and random phone call I got this week as well. Good timing. I’ll transcribe from memory.
Unknown woman caller: please stop calling me and then hanging up
me: um…. I haven’t been calling you.
uwc: yes you have, your keep calling me from this number and hanging up right away. You keep doing it, and I want it to stop.
me: it’s not me. [checks caller id on phone and phone history] I just checked my phone. I’ve never called this number, it’s not in my history.
me: [pause] I’m not sure who’s calling you but it’s not me
uwc: this is your phone number, so it’s your problem, you need to fix it.
me: maybe a computer is calling you, and somehow faking it’s coming from my number. [using my reasonable person at work voice] I’m really sorry but I really have never called you. I don’t know why you’re getting calls from my number. why don’t you block my number? that’ll fix it.
uwc: [now less angry] I don’t know how to do that.
uwc: hangs up
Given the anger level, I’m sure this was a real person. Not Google Duplex. I assume computers are pulling this stunt to discover which phone numbers reach live people.
Related, the New York Times ran a story this week Yes, It’s Bad. Robocalls, and Their Scams, Are Surging. Key stat is automated spam voice calls have “skyrocketed in recent years, reaching an estimated 3.4 billion in April.” With “an increase of almost 900 million a month compared with a year ago.” I’m sure you’ve noticed the uptick as well.
The economics are simple. If it’s cheap to try to scam everyone, then do it, and see who you catch. One of the most durable scams is the Nigerian Prince, who can’t get his inheritance money out of Nigeria, unless you send him a small sum immediately, after which he’ll split his fortune with you. This scam was first sent by snail mail, then moved to fax, then moved to email. In 2004, Bill Gates said “Two years from now, spam will be solved.” Which wasn’t quite right, but eventually AI got good enough at detecting email spam that it’s no longer a problem. But what AI giveth, AI taketh away. So please get ready for the dulcet tones of computer voice spam, with a completely realistic Nigerian Prince asking lonely non-tech savvy people about their health, and how they can split the prince’s inheritance money if they just send him $1000 bucks today. Preferably in bitcoin. This prince has time on his hands, so can talk to lonely people all day if that softens them up. As of right now, it’s not too late to call your parents to warn them. But don’t wait. In a couple of years, a computer sounding exactly like you will be asking them for their life savings. Over the phone.
2. Privacy is Dead. Some very good points made by Mims in his Wall Street Journal piece. Social networks include you, even if you aren’t on them. If someone you know joins a social network, they get asked to allow their contacts to get pulled. Not everyone does, but most do. And this gives them your name/phone number/email. For example if “1% of cellphones in London were compromised with malware, an attacker would be able to continuously track the location of more than half the city’s population.” And 270k people who used the Cambridge Analytica app got data for 87 million people. And while anonymizing that data seems like a good idea, “researchers with access to pools of anonymized data have found ways to identify individuals within it.” Mims concludes by arguing that everyone’s data is already out there, so it’ll be centralized either by governments (think China) or by huge tech companies (think Google, Facebook). Rather than say this is really really bad, Mims argues it’s more productive to say modern big data tech is here to stay. It has some good, and personally I think the good outweighs the bad. But opinions will very. Regardless let’s face the facts about the bad. And decide how to deal with it. The choice is government or regulated tech companies.
3. Graham Coop’s lab does the math. The golden state killer was found using publicly available DNA databases. Coop’s lab wrote up a good technical explainer on how that worked. Key sentences: “Under the assumptions we make here, it’s likely that a large percentage of people have at least one high-confidence genetic cousin in GEDmatch, and the number of 3rd-4th cousins found for DeAngelo—10 to 20—is not too far from the expectations.” Note the similarity here to Mims’ point just above about social networks. Both public DNA data sources and modern social networks can use the exponential power behind network math to identify you. Even if you’re not personally on these systems.
4. California rooftop solar required by law. This seems like a good idea. Who’s against solar? Well. If you think in terms of economics I believe this is a dumb idea. Solar rooftops are fine for volunteer private houses of course! But mandating them means more expensive and less new housing. And expensive housing has worse effects on people. And also (if you include the impact on housing costs) makes this an extremely overpriced way to cut carbon. I’m reassured that left/liberal Matthew Yglesias and right, or more correctly libertarian, Alex Tabarrok both hate the idea. Ygleasias There’s an easier way for California to build greener housing: just build more homes. Tabarrok Rooftop Solar is Expensive and Inefficient.
5. Chokers are the best people. Scott Sumner argues that if you do your best all the time in the NBA, you won’t get better in the playoffs. So you’re a choker. This is bad. But if you’re lazy most of the time, you will get better in the playoffs. So you’re clutch. This is good. He concludes by flip flopping conventional wisdom, saying chokers are (morally) the best people. Sumner is correct of course, and this is generally applicable.
6. Vest translates sounds into vibrations felt on skin, which after practice allows deaf to “hear” those vibrations. This is excellent. Quote:
That “unlocking” phenomenon, like adding a new sense, is hard to explain. How do a series of vibrations that supposedly reflect sound eventually have meaning when there’s no language assigned to them? How does the brain on the first day have no idea what a couple of vibrations on, say, the lower back means, but by the fifth day, know that they form a specific word? “My view is that the brain is a general-purpose computational device,” Eagleman told me. “You could take any kind of data stream and the brain will figure it out. I consider it the biggest miracle no one’s heard of.”
Note how generalized the brain is in it’s ability to pull in a brand new stream of sensory input.
7. Podcast on how big esports are. Esports are huge. Selling out stadiums for people to watch other people play video games live. What I liked about this podcast is it doesn’t assume you know much (that’s me), and walks you through how things got so big. Don’t listen if you know about esports already, but if you’re curious about why these are big and getting bigger, take a listen. link
Thanks for reading to the end. Have a good week.
Also, if you’re curious, here’s the original Yoda quote next to my updated one on the image at top:
The shroud of the dark side has fallen. Begun, the clone wars has.
The shroud of the computer talk has fallen. Begun, the spam wars has.
My weekly list of good reads. With commentary on why I liked them. This week has more than usual, so let’s go.
1. Robin Hanson, sex redistribution, viral outrage. To recap, Hanson wrote a short post arguing inequality in the distribution of sex may be similar in importance to inequality in the distribution of income. His quoted the news of a self-described involuntary celibate (incel) murdering 11 people by driving through a shopping district in Toronto. Hanson ended his post saying “Strikingly, there seems to be little overlap between those who express concern about income and sex inequality. Among our cultural elites, the first concern is high status, and the later concern low status. For example, the article above seems not at all sympathetic to sex inequality concerns.” Hanson’s post now has 500+ comments(!). It started going viral with this twitter thread asserting Hanson is motivated by evil. Then Jordan Weissmann in Slate asked Is Robin Hanson America’s Creepiest Economist? Which was predictably followed by someone claiming Hanson believes “women should fuck violent men“, complete with quote marks around those words. Sigh. Ross Douthat at the New York Times wrote a piece influenced by Hanson’s post on incels and sex robots, which of course was misinterpreted as well, leading Douthat to write a twitter explainer thread.
The viral dynamic here is simple and commonplace. Someone writes about topic X, which is associated with very bad thing Y. But the writer doesn’t talk about Y, or talks about Y casually, so people go bananas. For example, if X is genetics and IQ, then the very bad thing Y is racism. Sam Harris and Ezra Klein did this dance recently. Klein calls out Harris for talking X genetics/IQ, but not talking Y, saying “You did not discuss how race and racism act upon that outcome. You did not discuss it. I mean, amazingly to me, you all didn’t talk about slavery or segregation once.” (See my recent post on Harris/Klein here). James Demore’s firing from Google followed the pattern, where X was stats on sex differences and Y was sex discrimination. In Hanson’s case, X was involuntary celibacy/sex inequality and Y was rape/misogyny. Plus mowing down 11 people with a van. You’ll notice these examples have an author coming from a rather nerdy and analytic viewpoint. And the Y they ignore or treat causally is a white hot flash political point. A recipe for viral hate.
My two cents is if you write about an X that’s associated with angry topic Y, it’s best to explicitly comment on Y. Don’t ignore the elephant in the room. In Hanson’s case, it sounds like overkill, but he should have explicitly said rape and misogyny are very very very very very very bad. As is killing people by driving over them. Or just not used the incel example to make his point (often the safest approach if you’re very analytic and prone to misjudge how readers may react). I’m not saying this is logical. Or fair. Or will even protect you all that much. But it’s how the internet human nature works.
All that said, unlike some of the other examples I used above, I find Robin Hanson a brilliant and highly original thinker. I’m a fan of his books and his blog. And the people attacking him and his motives for his post on sex redistribution got it completely wrong. He’s obviously a very analytic academic trying to sort out the puzzles of human behavior, not encouraging or advocating for rape and misogyny. The claim would be utterly ridiculous if it weren’t now Hanson’s (permanent?) top google result. I’m glad at least his peers are supporting him: Alex Taborrok, Kevin Simler, Adam Ozimek, Garrett Jones.
Another pattern of the Two Minutes Hate I’ve noticed is these events create heat but no light. There’s little to learn about sex inequality from this dust up. On the other hand, I think it’s worth understanding the underlying pattern, which is what I’m taking a shot at here. If nothing else so others (like me) can avoid it. Another thing to learn from this go round is the psychology behind those involved, for which I highly recommend Scott Aaronson’s sympathetic (if at times chastising) post on what happened. Read it here.
2. Hominin in Philippines 700k years ago. Butchered rhino bones and stone tools were found on Philippines’s largest island, Luzon (paper). Homo sapiens evolved 300k years ago. So this means it’s Homo erectus, or an as yet unknown hominin. The analogy I’d draw is many people used to believe humans were unique in using tools, having culture, having a theory of mind. All of which have since been demonstrated in chimps and corvids. Similarly, some people believed Homo sapiens sprung rather discontinuously into being with language, arts, complex reasoning. But now we’re learning human behavior has far deeper roots in the homo lineage. Including, apparently, an ability to build rafts for sailing the seas. Best popular article I read is here.
3. Catching blood doping in marathons. Elite runners “tend to produce performances that vary from race to race by 1 to 1.4 percent”. When an athlete ran 4.22 faster than expected, he was caught. This seems a general method. It works best in timed sports like track or road racing, but if an athlete dramatically improves at any point in their career, it’s worth investigating. The author Alex Hutchinson is always good on the science behind distance running. link
4. Stories as the new social media format. “Stories” are short videos built of images, captions, overlays and effects. Snapchat invented them. Instagram and Faceook copied them. Newsworthy quote from Chris Cox, Facebook’s chief product officer: “The increase in the Stories format is on a path to surpass feeds as the primary way people share things with their friends sometime next year.” If you follow tech, worth knowing how incredibly popular Stories are, even if, like me, you are a wee bit too old to use them very often. link
5. Globalization in the information age. Richard Baldwin splits globalization into three eras. The first globalization (from the industrial revolution) is/was about moving physical goods. So moving textiles or iPhones. We’re talking David Ricardo. The second is/was about moving know-how. For example moving the knowledge for creating textiles or smartphones. The third will be about the virtual movement of people and human services. So we’re talking human services done via the internet, remotely across the globe. I found Richard Baldwin’s piece insightful and thought provoking (good), if a tad jargony/incoherent (bad). On balance I’d still recommend. link
6. Contingency of Marx’s popularity.Branko Milanovic argues Karl Marx’s popularity was contingent on three (not inevitable) turns of history. First, Engles spending two decades writing Das Capital. Second, the first world war thrusting Russia into chaos and the October Revolution. Lastly: “The fall of communist regimes brought it to its low point. But then –the third event—globalized capitalism that exhibits all the features that Marx so eloquently described in Das Capital, and the Global Financial Crisis, made his thought relevant again.” I’m far more skeptical of that last point compared to the first two. And if it were me I’d have emphasized Marx’s totalitarianism tendencies more. But the argument for the contingency of Marx’s historical influence is well made.
7. Credit Scoring using Digital Footprints. Link to paper, via Arnold Kling. Information “people leave online simply by accessing or registering on a website” is as good as FICO credit scores. Perhaps inevitable, and also perhaps inevitably regulated. No doubt this will blow up one way or the other in the next few years. Especially as Facebook and Google are the natural companies to do web behavior based credit scores, given their reach. Watch for it.
8. Marchettis constant of one hour of commute time per day. So 30 minutes each way. “For Marchetti, one hour is the basic limit for the total amount of travel that humans have been willing to put up with each day since the dawn of human society. He speculated that early humans travelling on foot at around 5km/h (3mph) would thus have a territory radius of 2.5km. To test the idea, he looked at the areas associated with individual villages in Greece – territories established over many centuries – and found that they tended to be roughly 5km across, supporting his claim.” Worth remembering as we enter an era of driverless cars. If you are working or sleeping during your driverless commute, does it count against the 30 minutes or not? link
9. Email are micromeetings. M.G. Siegler hates email, but now tries to write about that hate less often. Good for him. His recent post makes an analogy between hating meetings and hating email. Both are time sinks imposed upon us from outside. “Emails are basically little meetings that you get sucked into throughout a day. Some of them may take just a few seconds, others take far longer. Some may even take days or weeks.” Very much agree. Nails the dynamic. link
Here’s links and comments on what I enjoyed reading this week.
1. DNA joins the surveillance society. This week police arrested a 72 year old man accused of 51 rapes and 12 murders between 1974 and 1986. He was tracked down using DNA site GEDmatch. How this works is you get your DNA results from a major company like 23andMe or Ancestry.com. It requires a court order for police to get that data. But once you have your own DNA data, you can also upload it wherever you want. Including an open DNA site like GEDmatch. GEDmatch now has about 900k people on file. And it’s free to use by anyone. While you share 50% of your DNA with siblings, it drops to 3% with 2nd cousins. The matches found on GEDmatch were 3rd and 4th cousins, which investigators turned into a list of about 1000 people who met the right profile. They narrowed it from there. Here’s a good explainer.
The key takeaway is how inevitable this feels. It’s been discussed for years. But only now is open access DNA information getting complete enough that everyone with lots of 2nd and 3rd cousins living in the U.S. can (in theory) be tracked down. The two drivers here are health and plummeting DNA sequencing costs. And of course many countries are ahead of the U.S., with national DNA databases tied to healthcare. Related – see my post on how the surveillance society is a return to our gossipy forager past.
2. Voting for status. Was the 2016 election result driven by white economic anxiety, cultural status, or plain racism? That’s how the question often gets asked. Sometimes conflating those last two. Now the New York Times reports on a study saying it’s cultural status. Example: “While economic anxiety did not explain Mr. Trump’s appeal, Dr. Mutz found reason instead to credit those whose thinking changed in ways that reflected a growing sense of racial or global threat.”
Andrew Sullivan argues for cultural anxiety, driven by immigration. Quote:
Ta-Nehisi Coates has called these people witting enablers of white supremacy because they voted for Trump, conjuring up images of men in white hoods lynching and murdering African-Americans. But many of them voted for Obama twice. Clinton called half of Trump voters “a basketful of deplorables.” But a majority of white women voted for Trump. The left intelligentsia regards them as bigots, racists, xenophobes, and even “privileged” — attitudes and statements that are re-broadcast every hour of every day to the white and culturally anxious viewers of Fox News. What few on the left seem to see is that cultural anxiety, given the ethnic and cultural transformation of the last few decades, is an entirely predictable and entirely understandable response. If people felt that someone in charge actually saw their point of view, sympathized with it, and attempted even minor changes to accommodate it, we would have a different politics. But all they had was Trump. And all they still have is Trump.
It wasn’t their economic insecurity that gave us Brexit. It was that no one in charge even sensed their unease. Elites — and I count myself among the guilty — gave them nothing by way of reassurance or even a sense that they were understood instead of reviled. So all they had was Brexit. It wasn’t a rational decision; it was their only way to have their voices heard. Their pride and self-identity are bound up in it now, just as a critical slice of America’s is bound up in Trump. Which is why, despite the mounting evidence that the Brexit gambit is a disaster, they will never let it go.
Adam Ozimek argues back at Sullivan, pointing out that “there are very few immigrants in the U.S. counties that went strongest for Trump.” And “I think a better description of what we’ve seen is not that immigrants have drastically changed Trump country and they are rebelling, but that other types of decline have occurred -from economic to social- and the very few immigrants who live there are scapegoated.”
I think Ozimek is more correct than Sullivan. But the larger point this whole debate misses is voting is always about status. We don’t live in a malthusian society. Modern economic hardship is status, not starvation. Being rich is status. Being Harvard is status. Being coastal city is status. Being racist is status. It’s status status status all the way down. We need to start already knowing uneducated whites have lost status. Then we can debate how much is due to immigration versus jobs versus race versus education versus media portrayal. And perhaps if we work hard enough, we’ll do something about it.
3. Music streaming. Progress toward music streaming continues apace. From Mark Mulligan: “streaming is dragging the entire recorded music industry back into growth (all other sales formats are in decline). The recorded music industry is on track to become a streaming industry in all but name.”
4. Seagoing Neandertals. Quote: “A decade ago, when excavators claimed to have found stone tools on the Greek island of Crete dating back at least 130,000 years, other archaeologists were stunned—and skeptical. But since then, at that site and others, researchers have quietly built up a convincing case for Stone Age seafarers—and for the even more remarkable possibility that they were Neandertals, the extinct cousins of modern humans.” Seagoing Neandertals! Most excellent. My favorite bit of the week.
Here’s my favorite reads from this week, with comments.
1. Apple Customers love iPhone X, except for Siri. Ben Bajarin at Creative Strategies published an ungated version of their iPhone survey, and no surprise people love the new interface. That is, the replacement of the home button with a swipe based interface, and unlock using Face ID. I agree. I also agree Siri remains the weakest aspect of the iPhone. No surprise there either. Has a nice chart summarizing the survey results. link
2. Exponential performance improvements in tech behind machine learning expected for another decade. Termed Huang’s Law (not sure that name will stick), after Moore’s Law, much of machine learning runs on the GPU stack. Exponential performance improvements in that stack should continue for another decade, faster than Moore’s Law. This is an important trend. link
3. Physiological and Genetic Adaptations to Diving in Sea Nomads. Carl Zimmer writes up the new paper in Cell, on the genetic adaptations of the Bajau people. Quote:
As scientists peer deeper into our genes, they are discovering instances of human evolution in just the past few thousand years.
People in Tibet and Ethiopian highlands have adapted to living at high altitudes, for example. Cattle-herding people in East Africa and northern Europe have gained a mutation that helps them digest milk as adults.
On Thursday in the journal Cell, a team of researchers reported a new kind of adaptation — not to air or to food, but to the ocean. A group of sea-dwelling people in Southeast Asia have evolved into better divers.
It turns out the Bajau have large spleens, which helps them dive for longer periods of time. Possibly due to introgression from Denisovans(!). The idea that human evolution can occur on the thousand year time scale is accepted now more than before, due to breakthroughs in genomics and sampling ancient DNA. Greg Cochran puts it this way: “You should expect significant local adaptation in any population that is effectively endogamous and has fallen into a distinct niche for a reasonably long time, say a thousand years or more.” link
4. Land reform bill SB-827 in California dies in committee. This bill failing is bad (in my view), but the fact it even made it to committee is a kind of progress. To that question, Kevin Drum asks why progressives should force cities to get bigger. Matthew Yglesias responds, explaining why. It helps poor people move to rich cities, so they can earn more money. I agree. If you want a visual on how restrictive zoning is impacting cities, this post has some nice maps. In particular this one:
5. Hominins killed off large mammals. Starting with Homo erectus 1.8M years ago and continuing with modern humans, we’ve been driving the largest mammals to extinction. Paper. From Ed Yong’s writeup:
When hominins like Neanderthals, Denisovans, and modern humans spread through Europe and Asia, the average mass of mammals there halved. When Homo sapiens later entered Australia, the mammals there became 10 times smaller on average. And when they finally entered the Americas, with effective long-range weapons in hand, they downsized the mammals there to an even steeper degree. By around 15,000 years ago, the average mass of North America’s mammals had fallen from 216 pounds to just 17.
6. Kindle should have physical buttons. Craig Mod argues the Kindle device should have physical buttons, taking advantage of it’s dedicated use for reading. If you’re into hardware or software usability, definitely worth a read. link
And that’s all for this week! Thanks for stopping by.
Here’s comments and links on what I enjoyed reading this week.
1. Facebook Facebook Facebook. The 2018 deluge continues. Which aspects seemed underreported? I’d say: A. Only a vocal minority care about privacy. B. Hence privacy is weaponized in a proxy war for what people really care about — Facebook’s power over news and which political tribe wins. C. Regulations will likely backfire, locking in Facebook’s dominance. D. Facebook’s leadership has a natural cultural affinity to globalist, elite, college educated, status-quo power. E. Yet Facebook’s economic incentives derive from broad based advertising, making them natural allies to populist nationalism. Perhaps you noticed a tension between those last two. Continue reading Saturday Links 14-Apr-2018: Facebook, China Tech fusing with government, Birds see magnetic fields, Media a side hustle for tech giants→
1. End of Windows. An excellent play-by-play analysis by Ben Thompson on how Satya Nadella has shifted Microsoft away from it’s (massively successful) past, and towards it’s only realistic future. One where Windows is treated as a legacy business. link
Why? Let me attempt a (hopefully) constructive explainer. Thus I’ll crib the voxsplainer format for this post. Don’t be a hater.
Section 2. The Nerd and the Manager.
If you read the Klein/Harris published email exchange, you’ll note one person is arguing in what I’d call nerd-mode, the other in manager-mode. Everyone uses both modes. But for a civil discussion both people need use the same mode.
When Ezra Klein talks healthcare, he’s in nerd-mode. In the weeds. He made his name blogging about healthcare. He knows all the policy details. So for healthcare he can argue with experts, and critique them. But when Klein is talking race and IQ, he’s in manager-mode. Let me quote from his email to Harris:
Which brings me to the podcast. I really think that core discussion over the scientific dispute here is the important one, and I don’t want to present myself as the best person to have it. So to the extent I can persuade you that the disagreement is legitimate and good faith, I still think an actual expert in this field would be a better guest than me. The Heier note and Flynn piece only underscores the point: there are clearly experts on both sides of this, and I think there is something in the non-Murray side’s presentation you are having trouble hearing as serious, or as honest, and I think finding the boundaries of that disagreement would be the most interesting and enlightening conversation here. I am not a race and IQ expert and don’t play one on podcasts, so I don’t want to be the other side of that debate.
Manager-mode is about social proof. A manager knows how to listen to a team of experts and build a consensus. Or barring consensus, can select which expert(s) to follow. But a manager is not an expert. Klein says experts from the Vox piece disagree with the experts Harris cites. Harris replies:
As a point of comparison, you can see how Siddhartha Mukherjee handled Murray in his book The Gene, and in my most recent podcast with him. As I told Mukherjee, I don’t think he was fair to Murray, and I think he is bending too far in his definition of “intelligence,” but the discussion was far more respectful and balanced (and honest) than what you published in Vox.
Why not publish Haier’s rebuttal? His presentation of the science is far more mainstream that Nisbett’s (or Mukherjee’s, for that matter).
Harris’ is talking in nerd-mode. Harris knows the details. He knows which experts are right, which are wrong, which are spinning. And when you’re in nerd-mode (as I am right now writing this post), you find it nearly impossible to understand why anyone can disagree with your facts. You’re reluctantly led to conclude lingering disagreement must be due to ill motives. Trust me. I use twitter. I’ve spent plenty of time in nerd-mode hell.
But of course in real life human cognition is highly evolved to swim in the ocean of social proof. We mostly live in manager-mode. We it take for granted the views of our in-group are correct. This dynamic appears to me to explain what happened. Harris published the emails in nerd-mode. That backfired because most people read them in manager-mode, which was the mode Klein was using at the time.
This bodes ill for the upcoming Harris-Klein podcast. Harris will argue expert details. Klein will respond by pointing out to experts who disagree. But there may be a way out.
Section 3. Music break.
But first. All good Voxsplainers have a 1990’s music break. Here’s mine.
The awesomeness of Weezer shall not be denied. On to section 4.
Section 4. Framing the IQ debate around what’s changed since 1994.
The Bell Curve was written 24 years ago in 1994. Perhaps we can reconcile nerd-mode and manager-mode by looking at trends over time. The nerd can point to changing details. The manager can point to changing scientific/public consensus. Both can be happy.
By far the most popular critique of The Bell Curve when it came out was the New Yorker piece by Stephen J Gould. Take my word for it. I remember 1994. In fact I bought Weezer’s first album on a plastic disc in 1994 when it came out. It remains excellent. But I digress. Back to Gould. Scroll down to the Appendix below to read Gould’s original words (nerd-mode people only of course). Or just use my bullet point summary of Gould right here:
The claim that IQ and g can use a single number to measure something real in the brain is a fallacy. “Nothing in The Bell Curve angered me more”
Racial differences in IQ being mostly determined by genetic causes “is most surely fallacious”
Murray claims social stratification based on IQ will occur because meritocracy selects for it. This requires that IQ “must be depictable as a single number, capable of ranking people in linear order, genetically based, and effectively immutable.” And “The central argument of The Bell Curve fails because most of the premises are false”
Now compare this to the Vox piece authored by Eric Turkheimer, Kathryn Paige Harden, and Richard E. Nisbett. Call it the THN piece after their last initials. Fortunately they bullet pointed their argument as below:
Murray’s premises, which proceed in declining order of actual broad acceptance by the scientific community, go like this:
1) Intelligence, as measured by IQ tests, is a meaningful construct that describes differences in cognitive ability among humans.
2) Individual differences in intelligence are moderately heritable.
3) Racial groups differ in their mean scores on IQ tests.
4) Discoveries about genetic ancestry have validated commonly used racial groupings.
5) On the basis of points 1 through 4, it is natural to assume that the reasons for racial differences in IQ scores are themselves at least partly genetic.
Until you get to 5, none of the premises is completely incorrect. However, for each of them Murray’s characterization of the evidence is slanted in a direction that leads first to the social policies he endorses, and ultimately to his conclusions about race and IQ. We, and many other scientific psychologists, believe the evidence supports a different view of intelligence, heritability, and race.
For comparison, here is Harris’ summary of Murray’s thesis:
Human “general intelligence” is a scientifically valid concept.
IQ tests do a pretty good job of measuring it.
A person’s IQ is highly predictive of his/her success in life.
Mean IQ differs across populations (blacks < whites < Asians).
It isn’t known to what degree differences in IQ are genetically determined, but it seems safe to say that genes play a role (and also safe to say that environment does too).
What has 24 years wrought? I’d say Gould’s critique of IQ has fallen completely outside of acceptable scientific and social discourse. IQ is a meaningful construct is explicitly agreed to by THN in their 2018 critique of Murray. Gould’s two other points haven’t aged well either.
Now compare THN to Harris. The first 4 points from THN are similar to the first 4 from Harris. These are relatively secure. Though as noted by THN , experts differ on how strong they think those 4 claims are. But the 5th claim from THN and the 5th claim from Harris remain in public dispute. Namely, to what degree do IQ group difference have a genetic component, if any. This is disputed almost by definition since Turkheimer, Nisbett and Harden are in fact experts and they say it’s in dispute. So arguing in manager-mode, it makes perfect sense for Klein to justify his position. To add some additional social proof, here’s Kevin Drum saying the same thing in a recent short post. #5 is not mainstream public consensus.
One more point. In Klein’s 6000 word post on the allure of race science, he says:
Here is my view: Research shows measurable consequences on IQ and a host of other outcomes from the kind of violence and discrimination America inflicted for centuries against African Americans. In a vicious cycle, the consequences of that violence have pushed forward the underlying attitudes that allow discriminatory policies to flourish and justify the racially unequal world we’ve built.
To put this simply: You cannot discuss this topic without discussing its toxic past and the way that shapes our present.
I’d put it more bluntly. When you talk about IQ and race, you are providing (if misused) weapons of mass destruction to the worst elements of American society, both past and present. So if you decide to write about IQ, you can’t avoid talking about racism since anything you say can be weaponized. Which is why I needed to say so right here in my blogexplainer as well. Klein is correct about this.
Section 5. Where is IQ science going next.
Klein and Harris qualify as public intellectuals. But academics are the ones who define the acceptable limits of scientific discourse in their area of expertise. I follow Erik Turkheimer and Paige Harden on twitter, and it’s clear to me they wrote their article in a spirit of public service. And similarly, David Reich’s new book is public outreach as well. Academic outreach is a thankless task. Your colleagues claim you seek controversy to sell books. Partisans attack you. You can do permanent damage to your career.
It’s human nature to remember who your enemies are, even after you forget why. Read Ezra Klein’s excellent piece How politics makes us stupid, on Dan Kahan’s work. So even if our understanding of IQ continues to shift over time, I doubt existing reputations will change, for Murray or anyone else. There are always new (and old) reasons to dislike your out-group (1,2,3). [Update. Along those lines, Matthew Yglesias just wrote a post taking Murray to task not for his IQ position, but for his social policies, The Bell Curve is about policy. And it’s wrong.]
Appendix. Gould’s 1994 critique of The Bell Curve.
Here’s key quotes from Gould’s 1994 critique of The Bell Curve, plus a link to the full article. Here’s Gould:
The Bell Curve rests on two distinctly different but sequential arguments, which together encompass the classic corpus of biological determinism as a social philosophy. The first argument rehashes the tenets of social Darwinism as it was originally constituted……The theory arose from a paradox of egalitarianism: as long as people remain on top of the social heap by accident of a noble name or parental wealth, and as long as members of despised castes cannot rise no matter what their talents, social stratification will not reflect intellectual merit, and brilliance will be distributed across all classes; but when true equality of opportunity is attained smart people rise and the lower classes become rigid, retaining only the intellectually incompetent…..
The general claim is neither uninteresting nor illogical, but it does require the validity of four shaky premises, all asserted (but hardly discussed or defended) by Herrnstein and Murray. Intelligence, in their formulation, must be depictable as a single number, capable of ranking people in linear order, genetically based, and effectively immutable. If any of these premises are false, their entire argument collapses. For example, if all are true except immutability, then programs for early intervention in education might work to boost IQ permanently, just as a pair of eyeglasses may correct a genetic defect in vision. The central argument of The Bell Curve fails because most of the premises are false…..
Herrnstein and Murray’s second claim, the lightning rod for most commentary extends the argument for innate cognitive stratification to a claim that racial differences in IQ are mostly determined by genetic causes—small difference for Asian superiority over Caucasian, but large for Caucasians over people of African descent. This argument is as old as the study of race, and is most surely fallacious.
Nothing in The Bell Curve angered me more than the authors’ failure to supply any justification for their central claim, the sine qua non of their entire argument: that the number known as g, the celebrated “general factor” of intelligence, first identified by British psychologist Charles Spearman, in 1904, captures a real property in the head.
I closed my chapter in The Mismeasure of Man on the unreality of g and the fallacy of regarding intelligence as a single–scaled, innate thing in the head….