Style transfer

Frank Liu developed a technique he calls “style transfer” in which he renders an image in the style of another. 

Bhautik Joshi riffed off of that for video, resulting in experiments like this rendering of clips from Blade Runner in the style of Van Gogh's Starry Night. 

Prisma has capitalized on this technique by taking it mobile. I've enjoyed using the app to render some photos in my camera roll in a variety of styles. I had previously found expensive apps and plugins on the desktop that could do something like this, but now it's available in a free mobile app. Any sufficiently advanced technology is indistinguishable from a free app on your phone, eventually.

I'm looking forward to when Prisma can handle video, it's only a matter of time before style transfer videos like the one above are flooding your social media feeds.

A market opportunity for human ingenuity remains, however, for those who can actually transfer not just the visual style but the entire cinematic grammar of one artist to another. What if Werner Herzog directed Toy Story? What if Stanley Kubrick directed Star Wars? Who wants to see Terrence Malick's take on Captain America?

I draw much too much pleasure from style transfer in prose, like imitations of Hemingway, Cormac McCarthy, and the such. James Wood once wrote of McCarthy:

To read Cormac McCarthy is to enter a climate of frustration: a good day is so mysteriously followed by a bad one. McCarthy is a colossally gifted writer, certainly one of the greatest observers of landscape. He is also one of the great hams of American prose, who delights in producing a histrionic rhetoric that brilliantly ventriloquizes the King James Bible, Shakespearean and Jacobean tragedy, Melville, Conrad, and Faulkner.

Interview with Matthew Gentzkow

Due to this work, we now know that newspaper media slant is driven mostly by the preferences of readers, not newspaper owners. And by examining browser data, he discovered that people don’t largely live in internet “echo chambers”—that is, they don’t exclusively visit sites that align with their political bent. Product brand preferences, he found, are established early in life and endure long after exposure to essentially identical, less expensive alternatives.
 

That's from the introduction to this interview of 2014 Clark Medal winner Matthew Gentzkow. This immediately caught my eye because it echoes some ideas I have (which is perhaps ironic considering one of those points is about the over-estimation of Internet echo chambers).

I think the Internet has expanded, on balance, the volume of ideas on all sides that most people are exposed to, offsetting the echo chamber effect. What should concern us is how people have reacted to that broadened exposure; instead of pushing people to the center, it has increased polarization. That may say more about how we receive ideas that threaten our worldviews and tribal affiliations than it does about the inherent nature of the internet. 

Like Gentzkow, I also believe the reason so much advertising targets young people, even though it's the adults that have money, is to lock in consumer preference for life. In that respect much of that advertising is more efficient than it appears.

Frankly, this interview contains so much high quality material I can excerpt all day and still barely make a dent, so do read the whole thing.

Good news for parents who, on occasion, let their kids watch a bit of TV just to get a respite from care-taking duties.

This reflects what I think is an important conceptual point—that took a while to really sink in for us—which is that you can’t talk about the effect of TV without thinking about what it’s crowding out. TV viewing is shifting time around. And, really, for any new technology, any change that is shifting the allocation of time, its effect is the effect of that technology relative to whatever you would have been doing otherwise. 
 
That has pretty important implications for this question because if you think about children of different backgrounds and what else they might be doing with their time, it’s easy to imagine that for some kids, watching television is a much richer source of input than a lot of what it might be crowding out. TV has lots of language; it exposes them to lots of different people and ideas. 
 
It’s also easy to imagine kids for whom it could be a lot worse than whatever else they would have been doing. Educated, wealthy parents or parents with a lot of time to invest in their kids might be taking them to museums and doing math problems with them and so forth. I think part of the reason so many people writing about this assume TV is bad is that they themselves are in the latter group.
 

We have a strong norm in America about the corrupting influence of TV on children. I'm not sure how it arose or where it came from, but I'd love to know the history of that meme.

Regardless, what it means is that TV is often underrated for its positive aspects. I saw a paper once, though I can't seem to track it down, that showed that the introduction of TV in different countries and societies correlated with a strong rise in equality for a variety of groups including women and minorities.

That's not so surprising when you consider just how efficient television is at transmitting cultural norms. Humans love stories, and in this age those stories travel most efficiently to more people when encoded in the form of television and film narratives.

On the other hand, TV has had a negative effect on political turnout.

On the other hand, TV isn’t just political information; it’s also a lot of entertainment. And in that research, I found that what seemed to be true is that the more important effect of TV is to substitute for—crowd out—a lot of other media like newspapers and radio that on net had more political content. Although there was some political content on TV, it was much smaller, and particularly much smaller for local or state level politics, which obviously the national TV networks are not going to cover. 
 
So, we see that when television is introduced, indeed, voter turnout starts to decline. We can use this variation across different places and see that that sharp drop in voter turnout coincides with the timing of when TV came in.
 

This reminds me of an idea I've written about before, that in this age of near infinite content, we now gravitate towards an information diet that is much more reflective of our daily preferences than in the past. Newspapers of old started with the front page and included editorially prescribed sections in equal volume: World, Business, Sports, Entertainment, Autos, and so on.

I was always skeptical those sections merited equal surface area, but it wasn't until readers could actually consume anything they wanted that we had a true view of their preferences. The internet is perhaps history's great lab on consumer choice, and what it shows is that most people generally only want small doses of the main entree of hard news, and a lot more appetizers and dessert: sports, entertainment, celebrity gossip, clickbait self-help, pornography. 

That's why the addition of The Ringer is valuable for Medium. There are only so many tech confessional pieces even the most ardent tech enthusiast can handle in the Silicon Valley bubble chamber; scatter a few copies of US Weekly on the coffee table, and hang a flatscreen TV tuned into ESPN, and more people will visit more often.

My co-authors, Bart and JP, along with Sanjay Dhar, another co-author of theirs, had written a really important paper in the Journal of Political Economy a couple of years earlier that documented huge differences across U.S. cities in which brands are popular. They showed that that actually is correlated with the timing of which brands were introduced first in those cities, even though all of those introductions happened, for the most part, 50 or 100 years ago and few people remember a time when you couldn’t buy both. Say, for example, that we have two brands that have both been in a particular city for 50 years. If one was introduced 70 years ago and the other 50 years ago, you can predict that the one that’s been there for 70 years is going to have a much bigger market share.
 

We often think of first-mover advantage in sectors with network effects, perhaps none more clearly so than in messaging, with the odd geographically clustered favorites around the world. What Gentzkow notes here is that first-mover advantage can apply in consumer packaged goods, too. 

It's not that surprising, though I point it out for those who are always questioning why brands target unemployed millennials or kids without any income with advertising. Think about the loyalty fans have to sports teams from their childhood hometowns, long after they've moved elsewhere.

Our research results push back on that and say that, at least in this particular context, ownership is not really the key driver of slant and, in fact, a lot of the driver is actually coming from consumer demand. Not only does that say that you might not need to be as worried about ownership, but it also says that the welfare implications of this are a little more complicated because now consumers are getting what they want. 
 
We might think from a political, democratic point of view that it would be better if the public got different, more diverse information. But there’s going to be a welfare trade-off because we would be giving them content they would prefer less. If we want to give people diverse content that we think is good for democracy, then we have to get them to actually read, watch or consume it. And, you know, giving a bunch of people in conservative places some liberal newspaper—well, our results would suggest they’re not going to read it. So, that seems to have important implications for policy. 
 
But it comes with a really important caveat. The finding that ownership doesn’t matter in terms of a newspaper’s political slant is not a universal result. It doesn’t apply everywhere. It’s a statement about newspaper markets in the United States—a highly commercialized, relatively competitive setting, and a place where the political returns to manipulating the average content of a newspaper might not be all that big.
 

The chicken and egg question: did Fox News come along and satisfy a market need that conservatives weren't aware of, or did the market need summon Fox News out of nothingness?

If the filter bubble is not the internet's creation, but inherent to human nature, that argues for a much different solution than just exposing people to more ideas. Perhaps it's how the ideas are framed? How people are educated? Do we need to instill different mental models?

I'm fairly certain that taking an angry Trump supporter, cuffing them in a chair, locking their eyes open like Alex undergoing the Ludovico technique in A Clockwork Orange, and forcing them to watch Rachel Maddow for days on end isn't going to have the salutary effect one might suppose (and neither would force feeding a liberal Fox News).

The expanding Overton window

The ‘Overton window’ is a term from political science meaning the acceptable range of political thought in a culture at a given moment. It was the creation of Joseph Overton, a think-tank intellectual based in Michigan, who died in 2003 at 43 after a solo plane accident. His crucial insight, one which both emerged from and was central to the work of the think tank Right, was that the window of acceptability can be moved. An idea can start far outside the political mainstream – flat taxes, abolish the IRS, more guns in schools, building a beautiful wall and making Mexico pay – but once it has been stated and argued for, framed and restated, it becomes thinkable. It crosses over from the fringe of right-wing think-tankery to journalistic fellow-travellers; then it crosses over to the fringe of electoral politics; then it becomes a thing people start seriously advocating as a possible policy. The window has moved, and rough beasts come slouching through it to be born.
 

John Lanchester on the roots of Brexit, emphasis mine. The echoes in the rise of Trump in the U.S. are hard to miss.

Kipling asked a good question: ‘What do they know of England who only England know?’ But there’s a variation which, today, might be more relevant: ‘What do they know of the UK who only London know?’ The answer to both questions turns out to be the same: ‘Not nearly enough.’ England is so small, geographically, that it is easy to forget that it is also surprisingly big. There is no rich country of equivalent size that is more densely populated. The only country which has both more people than England and more people per square kilometre is Bangladesh. What this means, experientially, is that there is a kind of denseness to England and to Englishness; England is both very similar to itself and significantly different when you move ten miles down the road.
 

Related: this Clay Shirky Tweetstorm to his white liberal friends. Forget the internet filter bubble, if it even exists. Far more dangerous is the geographic filter bubble.

When geography is destiny, inequality is a given, but in certain countries the effect has amplified.

To be born in many places in Britain is to suffer an irreversible lifelong defeat – a truncation of opportunity, of education, of access to power, of life expectancy. The people who grow up in these places come from a cultural background which equipped them for reasonably well-paid manual labour, un- and semi- and skilled. Children left school as soon as they could and went to work in the same industries that had employed their parents. The academically able kids used to go to grammar school and be educated into the middle class. All that has now gone, the jobs and the grammar schools, and the vista instead is a landscape where there is often work – there are pockets of unemployment, but in general there’s no shortage of jobs and the labour force participation rate is the highest it has ever been, a full 15 points higher than in the US – but it’s unsatisfying, insecure and low-paid. This new work doesn’t do what the old work did: it doesn’t offer a sense of identity or community or self-worth. The word ‘precarious’ has as its underlying sense ‘depending on the favour of another person’. Somebody can take away the things you have whenever they feel like it. The precariat, as the new class is called, might not know the etymology, but it doesn’t need to: the reality is all too familiar.
 

This is amazing:

The white working class is correct to feel abandoned: it has been. No political party has anything to offer it except varying levels of benefits. The people in the rich parts of the country pay the taxes which support the poor parts. If I had to pick a single fact which has played no role in political discourse but which sums up the current position of the UK, it would be that most people in the UK receive more from the state, in direct cash transfers and in benefits such as health and education, than they contribute to it. The numbers are eerily similar to the referendum outcome: 48 per cent net contributors, 52 per cent net recipients. It’s a system bitterly resented both by the beneficiaries and by the suppliers of the largesse.
 

So much of communicating well is knowing your audience. Some of this is selection bias, of course, but Trump is the last man standing in the GOP because he is the voice of some large segment of America. When he preens and smirks, punctuating his misogyny and racism with finger jabs and thumbs up gestures, he wraps his hands around the id of the angry white working class and squeezes, like a cardiac surgeon pumping a human heart.

Electing one person, or three thousand?

Interview with the Wall Street Journal's Bret Stephens:

And if you’re obliged to make the choice upon pondering that, who would the choice be?

BS: Well, you’re asking me the same question twice. My answer is the same. The only person who counts in the administration is the president of the United States, Hugh. That’s the only person who counts. When George W. Bush decided to save the American position in Iraq by going against the advice of all of his wise men, of Jim Baker and the whole Iraq Study Group, and 90% of his administration, that was George W. Bush’s decision. So we have to bear in mind that this isn’t an administration we’re electing. It’s a person that we are electing. Who knows better than you what it means to have a commander-in-chief who lived his entire life, who lived throughout the entire Cold War, and doesn’t know what the nuclear triad is? It’s absolutely astonishing. And so it’s terrific to have Joe Dunford and you know, perhaps John Bolton and other people in positions of trust. But you have to have a president who bothered over the last 70 years to gain a cursory understanding of how the world works. And on so many issues, Hugh, on so many issues, I know not all of the issue, but on so many issues, this guy is just the antithesis of what I’d want a Republican president to be on foreign policy. When it comes to trade, when it comes to standing up to countries like North Korea, when it comes to standing up to guys like Vladimir Putin, Donald Trump is not a conservative. If you put…

HH: Bret, you don’t have to, I agree with you on all of that. I know the critique. Nevertheless, what about my argument that civilians owe people who are fighting the war the best of the two candidates for commander-in-chief. We don’t have the option to be conscientious objectors in the one part of the war that is part of our job, which is to pick a commander-in-chief.

BS: Listen, I think that for the United States, Hillary Clinton, as awful as I find her, is a survivable event. I’m not so sure about Donald Trump.
 

I've nothing to add here on Trump, but the idea of whether you're electing 3,000 people or 1 is a relevant question here as it is elsewhere (think about hiring a director, electing a CEO, etc.). Also worth asking if you're trying to hire the 1 person or one of the 3,000.

Romance, after the bloom

Ran across two good essays recently, both on later stage romance.

One, by Heather Havrilesky:

But once you’ve been married for a long time (my tenth anniversary is in a few months!), a whole new kind of romance takes over. It’s not the romance of rom-coms, which are predicated on the question of “Will he/she really love me (which seems impossible), or does he/she actually hate me (which seems far more likely and even a little more sporting)?" Long-married romance is not the romance of watching someone’s every move like a stalker, and wanting to lick his face but trying to restrain yourself. It’s not even the romance of “Whoa, you bought me flowers, you must REALLY love me!” or “Wow, look at us here, as the sun sets, your lips on mine, we REALLY ARE DOING THIS LOVE THING, RIGHT HERE.” That’s dating romance, newlywed romance. You’re still pinching yourself. You’re still fixated on whether it’s really happening. You’re still kind of sort of looking for proof. The little bits of proof bring the romance. The question of whether you’ll get the proof you require brings the romance. (The looking for proof also brings lots of fights, but that’s a subject for another day.)
 
After a decade of marriage, if things go well, you don’t need any more proof. What you have instead — and what I would argue is the most deeply romantic thing of all — is this palpable, reassuring sense that it’s okay to be a human being. Because until you feel absolutely sure that you won’t eventually be abandoned, it’s maybe not 100 percent clear that any other human mortal can tolerate another human mortal. The smells. The sounds. The repetitive fixations on the same dumb shit, over and over. Even as you develop a kind of a resigned glaze of oh, this again in, say, marital years one through five, you also feel faintly unnerved by your own terrible mortal humanness.
 

Or you should feel that way.
 

Another, by Alain de Botton:

Given that marrying the wrong person is about the single easiest and also costliest mistake any of us can make (and one which places an enormous burden on the state, employers and the next generation), it is extraordinary, and almost criminal, that the issue of marrying intelligently is not more systematically addressed at a national and personal level, as road safety or smoking are.
 
It’s all the sadder because in truth, the reasons why people make the wrong choices are easy to lay out and unsurprising in their structure. They tend to fall into some of the following basic categories.
 

Botton proposes a new form of marriage to follow on the previous two ages of marriage which he terms the marriage of reason and then the marriage of romance. He terms this the psychological marriage.

In the age of the marriage of reason, one might have considered the following criteria when marrying:
- who are their parents
- how much land do they have
- how culturally similar are they
 
In the Romantic age, one might have looked out for the following signs to determine rightness:
- one can’t stop thinking of a lover
- one is sexually obsessed
- one thinks they are amazing
- one longs to talk to them all the time
 
We need a new set of criteria. We should wonder:
- how are they mad
- how can one raise children with them
- how can one develop together
- how can one remain friends
 

Romance/marriage, as with many human institutions, is susceptible to human myopia. People are lousy at anticipating long-term consequences, and romance is particularly seductive with its immediate chemical rush. 

Acoustic pest detection

The U.S. Grain Inspection Service, Packers, and Stockyard Admininstration’s (GISPSA) standard quality assessment method involves sieving and visually inspecting a one kilogram sample: their guidelines “consider grains infested if the representative sample contains two or more live weevils, or one live weevil and one or more other live insects injurious to stored grain, or two or more live insects injurious to stored grain.”
 
However, since the larvae of many stored product pests grow inside grain kernels, where, Fleurat-Lessard notes, their “population density may be ten times more numerous than free-living adults,” a visually-inspected“clean” sample may actually be completely infested with rice weevil larvae. To look inside grains, laboratories use X-rays or resonance spectroscopy, but these techniques are too expensive and impractical to deploy in bulk grain lots.
 
But while rice weevil larvae are invisible, they are not inaudible: the “mean sound pressure” of rice weevil larvae feeding inside a wheat kernel is 23 dB, according to the USDA Agricultural Research Service. The idea, then, is that if you could somehow design sensitive-enough acoustic probes, combined with software to match the probes’ input against a database of field recordings, you might be able to monitor insect activity in stored grain automatically and detect infestations at the larval stage.
 

I had no clue such a thing as acoustic pest detection existed. Amazing.

Building a sound library of stored food insects was equally important – the field recordings on that Insect Noise in Stored Foodstuffs CD actually form the core of current acoustic pest detection databases. Years of research have gone into classifying the characteristic sonic signatures of different pest species at different stages in their lifecycles, to the point that a computer can now compare input from a grain silo’s acoustic sensor system against a library field recordings and tell you whether the rice weevil larvae eating your wheat kernels are sixteen or eighteen days old.
 

The smartphone is a form of human augmentation, the latest version of the “bicycle for the brain” metaphor from Steve Jobs or whoever it originated with. I'm looking forward to more sensory augmentation in compact form factors in my lifetime. The ability to increase the sensitivity of my hearing and have it plug into a database of sounds for enhanced recognition would open up a whole new world. Camping would never be the same again.

Process vs results in creative work

In the 1960s, a creative performance researcher named George Land conducted a study of 1,600 five-year-olds. Ninety-eight percent of the children scored in the “highly creative” range. Dr. Land re-tested each subject at five year increments. When the same children were 10 years old, only 30% scored in the highly creative range. This number dropped to 12% by age 15 and just 2% by age 25. As the children grew into adults they effectively had the creativity trained out of them. In the words of Dr. Land, “non-creative behavior is learned.”
 
Similar trends have been discovered by other researchers. For example, one study of 272,599 students found that although IQ scores have risen since 1990, creative thinking scores have decreased.
 
This is not to say that creativity is 100% learned. Genetics do play a role. According to psychology professor Barbara Kerr, “approximately 22% of the variance [in creativity] is due to the influence of genes.” This discovery was made by studying the differences in creative thinking between sets of twins.
 
All of this to say, claiming that “I’m just not the creative type” is a pretty weak excuse for avoiding creative thinking.
 

From James Clear.

If you've been following all the recent popular thinking about parenting, the recommendations that follow for improving your creativity won't be a surprise. As embodied in Carol Dweck's book Mindset: The New Psychology of Success or Paul Tough's How Children Succeed: Grit, Curiosity, and the Hidden Power of Character (how perfect is it that a book on grit is written by an author named Paul Tough, who sounds like a hard boiled detective), the vogue in self-help is sustained effort over innate talent, repeated practice over any one result.

If the path to creative breakthroughs is a probabilistic one, than an approach of sustained effort is more likely to yield success, and one's understanding that outcomes are not necessarily deterministic should render failure less psychologically crippling. A good combination for people, and entrepreneurs especially, is a small short-term memory, so you can move on from your mistakes, and an ample long-term one, so you can learn from them.

I learned the hard way as an undergrad the danger of waiting for inspiration when it came to my creative writing classes, or even just any form of writing which might be regarded as “creative” work. That's just a recipe for having a deadline scare you across the finish line in a frenzied late night of work, and it's hard to generate great work under such stress or fatigue, even if it's a great motivator.

I've come to see the simple act of writing here on my blog as a form of meditation. At some point it became a habit, one I've felt a true guilt for neglecting these past few months while I try to ramp up at a new job, and while it can seem like a maddening and arbitrary task I hold myself to, if even for just 15 minutes a day, the process bears long-term fruit. Writing forces me to clarify my thinking, and getting my thoughts out of my head and into a written piece frees CPU cycles in my head. The inspiration comes in the perspiration, as writers say, even if there usually isn't much sweat involved.

The secondary effects of writing have increased in this age, too. Readers have easier ways of finding and responding to my work, and that interaction has been not only intellectually rewarding but socially and professionally, too.

The art of the auction

Of all his crafty sales tactics, the one used to greatest effect by Christie’s 52-year-old global president – the auctioneer on top of the wave of wealth in the world’s art market – is the query: “Are you sure?” He asks it as one of the last two bidders in an auction drops out, threatening to finish it. As he halts the action for a few seconds – what he calls “the auctioneer’s pause” – pressure builds on the reluctant bidder.
 
Perhaps he will deploy it on Monday, as he stands in Christie’s saleroom in New York and attempts to sell a 1917 Modigliani nude for more than $100m at the start of what Christie’s and its biggest rival Sotheby’s hope will be another record-setting week of art auctions. 
 
It is a deft inquiry for it is impossible to dismiss – of course the bidder is unsure. No one is certain because no one knows what a work of art is worth – perhaps a few dollars more, or a few million.
 

Short, fun profile of Christie's top auctioneer Jussi Pylkkanen.

Auctions are unusual in the 21st century – most things, even luxury items such as watches and clothes, sell at fixed prices although there is some room to haggle. Even many items on eBay, the electronic platform, are sold at fixed prices. In The Dynamics of Auction, Christian Heath, a professor of work at King’s College, London, describes them as “a somewhat anachronistic method of selling goods, more common perhaps to traditional agrarian societies than post-industrial capitalism.”
 
They are still used for art because every painting is different and has no intrinsic value – it does not yield anything and the cost of manufacture is usually tiny. They are also a good way to get high prices – when buyers compete against a deadline, they behave differently. The desire not only to acquire it but to beat others causes what Deepak Malhotra, a Harvard professor, terms the “emotional arousal” of auctions.