Open Thread, 1/3/2016

A friend of mine, a man-in-tech of eminently WASP background of moderately liberal orientation in case you care, has been bemoaning the downstream consequences of the floundering of Marissa Mayer of Yahoo!, the confused direction of 23andMe under Anne Wojcicki, and finally, there is Elizabeth Holmes and Theranos. These are three separate cases. I don’t believe that Mayer ever had a high probability of success at Yahoo!; it was a legacy of the late 1990s that had never managed to pivot and find a new direction. Even a company with as many resources as Microsoft has been having difficulties finding domains where it can both be dominant and produce growth. Yahoo! never had the success of Microsoft, and so doesn’t have as much margin.

Of course long odds does not mean that a turnaround was impossible. I’m old enough to remember when Steve Jobs came back to Apple and many people shrugged, as it looked inevitable that the firm’s future was in diminishing returns on the margins of the consumer and educational PC market in the shadow of Microsoft and PCs (at one point Microsoft was making more profits on sales of Apple computers than Apple because of the Office productivity suite from what I recall). Obviously it didn’t play out like that. If Marissa Mayer had turned around Yahoo! she’d be a genius. If she failed, as she seems likely to at this point, there will be some sort of exit strategy where she will save face, and there will be no threat to her solid position among the firmament of American oligarchs. The likes of Mayer are seeking glory like the Roman Senators of yore. Only a few will go down in history as men or women of renown, but all will die rich and comfortable.

As Marissa Mayer is a beautiful and intelligent woman who will reproduce above replacement, of course I was always rooting for her. Beauty and intelligence are good. In some measure ambition is as well. And when was being coldly analytic an insult? I couldn’t care less about Yahoo!, as it’s irrelevant to the American economy as a whole. But I cared a bit about Mayer’s success, though I was always pessimistic. Probably some of the same dynamic applied to my attitudes toward Elizabeth Holmes, who seemed smart and attractive, though I didn’t pay attention closely to her claims or business. The issue which many, including Steve Sailer, have suggested, and which my friend agrees with, is that because of the cultural expectation that the tech sector promote “diversity”, Holmes went somewhat under the radar due to dampened due diligence. An extreme case of this sort of thing occurred with Jayson Blair at The New York Times, where Howell Raines admitted that as a liberal white Southerner he was inclined to treating Blair “gently”. Ballooning of reputations and media hagiographies of tech visionaries are not without precedent (it’s a substantial proportion of the copy of glossy business magazines), but some have wondered if Elizabeth Holmes’ story was just too good to inquire too closely. Not only did she fit the stereotypical motif of Silicon Valley entrepreneurs (e.g., Stanford dropout attempting to “disrupt” a sector with innovation), but she also served as an exemplar of a young woman who falsified the idea that the sector was doomed to remain a man’s game.

With 23andMe and Anne Wojcicki, I have no idea what’s going on, but it is, and always has been, confused. I found out from inside sources what happened with the initial FDA letter, and what I can say is that it wasn’t part of the plan in any way but a total f**k up. 23andMe has a high valuation, and an incredible database, and seems totally pivoting to the health market, shaking off its past in recreational genetics (read: genealogy, etc.). But from what I’ve heard Ancestry has surpassed, or is close to surpassing, 23andMe’s database (speaking of, as of this writing Ancestry has a 20% of sale, on checkout at Amazon). I wonder if 23andMe isn’t getting a bit overconfident, as Ancestry is going to shift into the health space too.

Obviously biases are real. CEOs tend to look a certain way. They’re far taller than average. And in Silicon Valley they’re disproportionately white males, as in the rest of corporate America (though less so). The die is probably loaded in favor of white males in relation to getting to the top of management. I’ve had enough experience in “industry” (i.e., the real world), as they would say in academia, to know that “corporate culture” often does have connotations which exclude particular groups naturally. If, for example, you are a business person in South Korea, the marathon drinking sessions are going to disadvantage many women and teetotalers. I think one of the reasons that Asian Americans, and in particular East Asians, are under-represented in management roles in Silicon Valley in relation to their representation in engineering positions has to do with personality, cultural norms (e.g., Asian parents not emphasizing sports and the sort of comradeship that it engenders and translates into the business world), and just the “look” (too many Asian American engineers are short and not fit).*

But I think the example of Elizabeth Holmes suggests that shifting the playing field a bit in business journalism does no one a service. The next time a young blonde attractive woman makes a pitch to investors no doubt one prior that is going to rattle in their brains is going to be “is she going to be another Elizabeth Holmes?” That’s just a cognitive bias. Though privately people will likely state this openly.

Speaking of cognitive biases, I’m about halfway through The Secret of Our Success: How Culture Is Driving Human Evolution, Domesticating Our Species, and Making Us Smarter. Yes, it’s as good as readers have told me. But, I’ll be honest and state that I knew most of it already in the generality, though not the ethnographic details (for example, I did not know that Inuit ate deer feces like “berries”; thanks for that Joe!). Part of this has to do with the fact that I’ve kept up on the author’s research since 2004, when I encountered his models of skill decay in Tasmanian Aboriginals. And, I’m familiar with the fields of cultural evolution more broadly. Additionally, I’ll bring it up in my review, but I think that he wrote at an unfortunate time when it comes to drawing lessons from human genomics, because some of his assertions have been falsified! (he wrote the preface in January of 2015, a year ago).

Reading The Secret of Our Success after I finished Consciousness and the Brain was also pretty lucky, as some of the assertions made in the former book seem much stronger and robust after the background provided by the latter. In the near future I plan to re-read both Origin of Mind: Evolution of Brain, Cognition, and General Intelligence and The g Factor: The Science of Mental Ability. Sometimes re-reading is about refreshing, but I don’t feel that in this case I understood human evolution well at all when I read these previously.

Steve Pinker has something out in The Wall Street Journal, Steven Pinker on New Advances in Behavioral Genetics:

But new studies looking for small effects of thousands of genes in large samples have pinpointed a few genetic loci that each accounts for a fraction of an IQ point. More studies are in the pipeline and will link those genes to brain development, showing that they are not statistical curiosities. The emerging picture is that most behavioral traits are affected by many, many genes, each accounting for a tiny percentage of the variance.

When Pinker says that studies are in the pipeline, he’s thinking of concrete studies which are in review, and which will blow your minds (though the usual suspects will obfuscate and ignore).

In relation to Steven Pinker, he gets a lot of hate and dismissal. Even Joe Henrich in The Secret of Our Success uses him as somewhat of a foil, though in a good-natured manner (Henrich and Pinker are now colleagues at Harvard, so collegiality is probably for the best). I still think that for writing The Blank Slate alone Steven Pinker will go down in history as an important thinker (though The Language Instinct was the most revelatory book of Pinker’s for me).

When I run TreeMix I often get gene flow edges from Africans or to Africans from a region of the graph basal in eastern Eurasians. I ignore these because they don’t make sense. But they don’t make sense because I’m missing something in the bigger picture. I’m sure a year or two years or three years from now it will all make sense. Just filing this away as results which I can’t make heads or tails of, but which are telling us something with Delphic clarity.

On to genetic data sets. I’ll post on this soon. But the 1000 Genomes are a disparate bunch when it comes to South Asians. The Tamils and Telegu speakers have three Brahmins each in them. Also, both groups have rather endogamous low caste populations as a small subset, distinct from the broader mass. The Gujaratis are highly structured, with a large cluster of Patels, but also various other groups in the mix in a cline out toward Northwest Indians (so I assume middle castes and Guju Brahmins). The Punjabis, sampled from Lahore, are also strange, because they exhibit a very extreme cline, from near the Gujaratis all the way toward West Asian/European populations as far as Pathans. Does this have something to do with people of Muhajir background or mix identifying as Punjabi? Finally, the Bengalis are curious, because they are different from the other South Asian groups in exhibiting minimal structure. These were people sampled in Dhaka, but the cluster is very tight, except five individuals who are closer to the South Indian groups. Two of these five have sample numbers adjacent, so I wondered if they were collected together. Unlike the other Bengalis these individuals don’t seem to have substantial East Asian admixture. I have no idea what group this might be, but I have a hunch that they’re derived from an endogamous caste (probably Hindus) who migrated to the area of Bangladesh in the last few hundred years from another region of South Asia. Finally, I have to note that the Bengali populations exhibit far fewer individuals with long runs of homozygosity than the other South Asian groups. Less than the Punjabis or South Indians, which stands to reason since these groups engage in high levels of consanguinity. But also lower than Gujaratis. The implications of this later….

Xinjiang Seethes Under Chinese Crackdown. I wish the media would explore the relationship of Uighurs, and Hui Muslims, in Xinjiang (the latter are called Dungans in Central Asia). That would get at the ethnic vs. religion tensions. Though in China proper the Hui are often proud of their Islamic identity, in Turkestan their affinities to the Han group, both physically and linguistically, become salient. Because of their Muslim religion and martial character the Hui/Dungans were often used as enforcers of Chinese hegemony by the Manchus. A somewhat greater number of Muslims in China are Hui than Uighur.

I stole my kids’ candy-canes. It was for their own good! Am I a bad person? Christmas was fun. My wife read Selfish Reasons to Have More Kids, but I did not (aside from a chapter Bryan Caplan asked me to read to check for any issues). I don’t need selfish reasons to have more kids. I like kids. Christmas was fun when I was a kid. But it’s more fun by far when you have kids. Watching my ~1.5 year old son sit down with his new board book is pretty awesome. The world is his oyster. Or at least toy truck.

I haven’t read the GCTA isn’t all that paper in PNAS. Yes, it’s not titled that, and also, am I a bad person? Yaniv Erlich is a good person, as he has read and responded. The Twitter reaction seems to be skeptical, but cautious. The reality is when the great Alkes Price or Michael Goddard weighs in we’ll know if this is simply a pretender to the throne.

A big deal at some point: phenopredict21.

Uncovering the genetic architectures of quantitative traits. On my “to-read” after the GCTA-sucks papers. (if you don’t know what GCTA is, shame on you! Read this: GCTA: A Tool for Genome-wide Complex Trait Analysis).

Edge, 2016 : WHAT DO YOU CONSIDER THE MOST INTERESTING RECENT [SCIENTIFIC] NEWS? WHAT MAKES IT IMPORTANT? To some extent if it doesn’t match =~ /CRISPR/ somewhere I’m underwhelmed. Did you see In vivo genome editing improves muscle function in a mouse model of Duchenne muscular dystrophy?

Michael Shermer: Murdering the facts about Homo naledi? Shermer also parrots those who criticize the H. naledi team for not publishing in Nature or Science. I happen to have been in Washington, D.C., when Lee Berger gave a private presentation on the findings over two years ago. And he already told me that the key was getting this research out quickly, and, dumping the data out there so that others could have access to it. It’s not just about H. naledi, it’s about changing how palaeoanthropology is done.

For those of you in Houston, I know that Cooking Girl has better Yelp reviews than Mala. Doesn’t deserve it in my opinion, though I’ll have to sample Cooking Girl more than once….

Also, I’ve spent some time in Austin, TX, recently. I lived in Portland, OR, years ago. Those who say that Austin is like the Texas version of Portland seem to totally capture it.

* I say Asian Americans, because internationals have obvious cultural handicaps in an American corporation.