How Computers Might Conquer the Game of Go

MIT Technology Review writes,

thanks to the work of Christopher Clark and Amos Storkey at the University of Edinburgh in Scotland. These guys have applied the same machine learning techniques that have transformed face recognition algorithms to the problem of finding the next move in a game of Go. And the results leave little hope that humans will continue to dominate this game.

Pointer from Tyler Cowen. As I read the article, they have been following a strategy very similar to what I proposed six months ago.

If they are as close to success as the article indicates, then the world of Go is about to be completely upended. With Othello or Chess, most of what is knowable had already been articulated by humans by the time that computers came along. At least as far as Othello is concerned, computers did not come up with any new strategy or tactics. They just got more skillful than humans at making the best choice in close situations. With Go, my guess is that there may still be a lot left to be discovered about the game. If so, then computers will soon be in a position to make the discoveries. Even if there is nothing new to be discovered, once computers start making fewer mistakes than humans, human Go players will soon be studying computer games.

The Two Languages of Libertarians

Daniel Klein classifies libertarians into challengers and bargainers. (He has a third category, “royalty,” to describe Milton Friedman and Adam Smith, who managed to achieve very high status.) Klein uses as an example of a topic the minimum wage. A challenger is someone who will say that the minimum wage should be abolished, while a bargainer is someone who sill say that the minimum wage should not be raised. Pointer from Tyler Cowen, who Klein pegs as a bargainer (I would agree).

Some comments:

1. I would describe challengers and bargainers in terms of language. Challengers use the language of certainty. “This is what I think, and people who disagree are just wrong.” Bargainers use the language of doubt or compromise. “Here is where my opponents and I agree, and here is where I think they are mistaken.”

2. I am mostly a bargainer. However, when I write posts using challenger language, I get a lot more praise and mention among libertarians. In fact, I have tried to keep myself from being influenced by such reinforcement.

3. You might be able to adapt these linguistic differences to other parts of the political spectrum. For example, I imagine that Paul Krugman evolved into the writer he is because he could not resist the positive reinforcement he received for expressing anger and certainty.

4. I think that Klein’s disctinction explains why I prefer having my own blog. I think it is fair to describe Bryan Caplan as more of a challenger, and when we were both at EconLog our styles clashed.*

5. Klein is never clear on whether the he is drawing an intellectual distinction between bargaining and challenging or whether he is making sociological observations. In fact, most of the talk strikes me as observations about differences between challengers and bargainers in terms of their personalities and social circumstances. For example, he says that the challengers tend to draw cult-like followings. On the other hand, he does say that an individual can make a choice about which stance to adopt, and it may even be possible to adopt different stances in different circumstances. That makes it sound more like an intellectual distinction. Bargainer that I am, I am trying to split the difference between making an intellectual distinction and making a sociological distinction, so that I want to emphasize linguistic differences.

6. I think that one difference, which can be viewed as intellectual but is probably grounded in personality, is that of certainty vs. doubt. The libertarians who Klein classifies as challengers strike me as highly certain. The bargainers have doubts. For example, challengers are quite certain that the world would be a better place with open borders, if drug laws were abolished, and so on. As a bargainer, I think that this is likely to be the case, but I am not so all-fired certain. Since challengers do not give much thought to being wrong, the fact that they are in a minority on an issue never bothers them. When I am in the minority, I question my own position–although I try to question my own position in any event.

7. In terms of what Jeffrey Friedman calls “the libertarian straddle,” challengers rely more on the philosophical a priori case for liberty. Think of Rothbard and the non-aggression principle. Bargainers rely more on the empirical economic case for liberty, which is that societies with more economic liberty tend to be more prosperous.

*This is all hindsight, in that I left EconLog primarily to pursue an ed-tech start-up. That did not go well, although I did learn a lot about how software had changed in the 15 years since I had been out of it.

Innovation Predictions

Nouriel Roubini writes,

A patient in New York or London may have his MRI sent digitally to, say, Bangalore, where a highly skilled radiologist reads the scan. However, that highly skilled radiologist in Bangalore may only be paid a quarter of what a New York radiologist would earn for reading tests.

It raises the question: how long before a computer can read those images faster, better, and cheaper than that Bangalore radiologist can?

Pointer from Tyler Cowen.

That sounds like a fair point. In general, however, I think that the forecasts for game-changing innovation made by Roubini and others are too aggressive. I do not share his enthusiasm for MOOCs, as you know.

For another bullish-on-robots, bearish-on-humans take, consider William H. Davidow and Michael S. Malone:

If you doubt the march of worker-replacing technology, look at Foxconn, the world’s largest contract manufacturer. It employs more than one million workers in China. In 2011, the company installed 10,000 robots, called Foxbots. Today, the company is installing them at a rate of 30,000 per year. Each robot costs about $20,000 and is used to perform routine jobs such as spraying, welding, and assembly. On June 26, 2013, Terry Gou, Foxconn’s CEO, told his annual meeting that “We have over one million workers. In the future we will add one million robotic workers.” This means, of course, that the company will avoid hiring those next million human workers.

Read the whole thing.

Frankly, I think that the biggest game-changer over the next fifteen years will be virtual/augmented reality that makes meetings among people from remote locations effective. If it comes off, it will reduce the significance of innovations in transportation, such as self-driving cars. It will also provide a platform for higher productivity in the New Commanding Heights of health care and education.

Here, let me make some predictions of when innovations will be well established (meaning that they have changed everyday life for many people), and I hope I do not err by being too aggressive.

Year Innovation
2020 Computer diagnosis based on lab results and other data
2025 Virtual/augmented reality enables people in remote locations to have meetings that feel “live”
2030 Food manufactured using bio-engineering rather than slaughtered or harvested
2040 Cures for all major diseases except cancer
2040 Personalized, computer-based education instead of classrooms
2045 Cure for cancer
? Fossil fuels overtaken as energy source by solar and/or nuclear power
? Medicines or implants that ensure high intelligence and conscientiousness
?? Drexler’s vision for nanotechnology, driving the cost of physical goods to near zero

As for the issue of human obsolescence, I do think that we will see a trend toward more and more leisure. This will raise all sorts of questions of who deserves to have what provided for them. Right now, we say that people aged 67 or so deserve Social Security and Medicare. And people who can command only low wages (already obsolete in some sense?) deserve Medicaid and food stamps. And kids who can get in deserve the leisure aspects of college. My guess is that we will struggle quite a bit over the next forty years to adapt the social bargain concerning leisure.

Smart Grid Problems

From the NYT,

Although the goal is to shift consumption to off-peak hours when cheaper, cleaner electricity is available, experts say it is still many years away, despite billions in federal subsidies that have helped finance the switch to the so-called smart grid.

Pointer from Tyler Cowen.

The article ends with a quote from an official in Maryland’s Office of People’s Counsel.

“I’ve never seen an analysis that shows that shifting my dishwashing, clothes-washing and clothes-drying load is going to make a significant impact on my monthly bill,” he said. “It’s just not that much electricity.”

I do not find this statement persuasive. Perhaps he needs to be charged more for use of electricity at peak times in order to get himm to understand the concept of dynamic pricing.

eP*/P

Those are the symbols for the “real exchange rate,” or the terms of trade, both measured inversely, or “competitiveness,” measured directly. That is, when this expression goes up, the real exchange rate depreciates, the terms of trade worsen, and competitiveness improves. e is the nominal exchange rate. Say that we are Japan, and e is yen/dollar. As e goes up, it means that our exchange rate is depreciating. (I am forever confused by that way of writing e, but that’s how it’s done.) P* is the domestic price index of our trading partners, and P is our domestic price index. Suppose that we (Japan) are relatively deflationary, which means that P*/P is going up. That has the same effect as a currency depreciation.

It appears to me that Japan is experiencing both a nominal currency depreciation (a rise in e) and an increase P*/P. That means that Japan is certainly experiencing a real exchange rate depreciation, or a real deterioration in its terms of trade. It has to give up more Toyotas to import the same amount of beef. In terms of purchasing power in world markets, the Japanese are becoming worse off.

At the same time, Japan has become more competitive. Japanese consumers will be inclined to import less beef. Toyota will find itself able to export more cars.

The question I have is this: when does Japan succeed in inflating away some of its debt? In terms of world purchasing power, it is already doing so. Japanese holders of government bonds are earning negative returns relative to the cost of a consumption basket. But that does not help the government. The government needs an increase in yen-denominated tax revenue.

Possibly related: Brad DeLong writes,

That process–the rise in domestic nominal prices and wages, and the larger fall in the nominal value of the currency–may derange the price system and so disrupt aggregate supply. The new equilibrium may be one in which the real depreciation of the currency is expansionary in the sense that it tends to push real aggregate demand above potential output. But the economy may nevertheless be in depression, if the process of getting to the new equilibrium has entailed nominal price swings large enough to have been sufficiently disruptive to the market-mediated division of labor. Weimar 1923.

Pointer from Mark Thoma.

I see that as the crux of the issue. If investors lose confidence in Japan’s bonds, the Japanese government loses its ability to borrow. When you lose the ability to borrow and you are running large deficits, watch out.

UPDATE: Read Tyler Cowen’s post on this topic.

Attitudes Toward Risk and Growth

Michael Hanlon writes,

Could it be that the missing part of the jigsaw is our attitude towards risk? Nothing ventured, nothing gained, as the saying goes. Many of the achievements of the Golden Quarter just wouldn’t be attempted now. The assault on smallpox, spearheaded by a worldwide vaccination campaign, probably killed several thousand people, though it saved tens of millions more. In the 1960s, new medicines were rushed to market. Not all of them worked and a few (thalidomide) had disastrous consequences. But the overall result was a medical boom that brought huge benefits to millions. Today, this is impossible.

Pointer from Tyler Cowen.

I think that the problem goes beyond rational risk aversion. One of the findings in behavioral economics is that people exhibit loss aversion. That is, they will avoid rational risks because they regret losses much more than they enjoy gains. It seems probably to me that government agencies exhibit at least as much loss aversion as do individuals.

Hurricane Katrina and World War II

Tatyana Deryugina, Laura Kawano, and Steven Levitt write,

Four years later, Hurricane Katrina victims are less likely to be unemployed.

…If moving costs (either financial or psychological) are high, then people will rationally forego higher earnings available elsewhere unless the expected benefit of moving is large enough to outweigh the fixed cost. The forced relocation caused by
the hurricane required displaced residents to pay the moving costs, leading to higher wages (although potentially lower utility levels).

Pointer from Tyler Cowen.

The PSST interpretation would be that forced relocation helps the economy more quickly find patterns of sustainable specialization and trade. That is my interpretation of the role that the second world war played in getting the United States out of the Great Depression. Millions of men were uprooted, and many of them chose to relocate to locations with better opportunities. When my late father attended a reunion of Soldan High School in St. Louis, he was stunned by the distance that many of his former classmates had traveled, both geographically and economically, from their home town. He reflected that this never would have happened without the war.

Standard macro would predict (and did predict) a major recession following the war. So might PSST, given the challenge of transition from wartime to peacetime. However, the fact that many returning servicemen actually thought about where they wanted to live after the war helped make the transition work. The paper on Hurricane Katrina suggests a similar effect.

The Sociology of Economists

Marion Fourcade, Etienne Ollion, and Yann Algan write,

we document the pronounced hierarchy that exists within the discipline, especially in comparison with other social sciences. The authority exerted by the field’s most powerful players, which fosters both intellectual cohesiveness and the active management of the discipline’s internal affairs, has few equivalents elsewhere.

Pointer from Tyler Cowen. As I describe in my macro memoir, Stanley Fischer, now vice-chairman of the Fed, controls a remarkable proportion of the sub-discipline of macroeconomics. For better or worse–and I strongly believe it is for worse–he has decided who is a macroeconomist and who isn’t.

Paul Krugman’s take:

It has been all too obvious that there are people with big reputations who can push equations around but don’t seem to have any sense of what the equations mean.

My only quarrel with that statement is that I believe it applies to equation-pushers from saltwater institutions as well as those from freshwater institutions.

Is Demography (Economic) Destiny?

The Economist blog writes,

An ageing population could hold down growth and interest rates through several channels. The most direct is through the supply of labour. An economy’s potential output depends on the number of workers and their productivity. In both Germany and Japan, the working-age population has been shrinking for more than a decade, and the rate of decline will accelerate in coming years (see chart). Britain’s potential workforce will stop growing in coming decades; America’s will grow at barely a third of the 0.9% rate that prevailed from 2000 to 2013.

Pointer from Tyler Cowen.

Along seemingly similar lines, Karl Smith writes,

It’s no accident that this phenomenon appeared in Japan first. As its population began to stagnate well before the rest of the industrialized world, investors found themselves with loads of capital, a dearth of workers, and repayment terms they could not meet.

First, think about this in the absence of inter-generational transfer schemes like Social Security.

1. If people live longer than they used to, then they either have to produce more (probably by retiring later) or consume less.

2. If birth rates decline, then you let capital depreciate faster than it would otherwise. Think of an economy where the only capital goods are houses that stay in good condition for fifty years. When birth rates are rising, you need to keep using some houses longer than fifty years, even though they no longer are in good condition. When birth rates are falling, you can take some houses out of service before fifty years, even though they still are in good condition.

This seems quite straightforward to me, and it is does not suggest that demographic changes should be highly disruptive. I am not persuaded by just-so stories about Japan. One can conjure many such stories. For example, maybe Japan slowed down because its corporatist approach to capital allocation was only effective for a decade or two.

The Year of Flawed Books

Writing a “best books of the year” post for 2014 means choosing among flawed books.

Six months after Piketty’s Capital made its splash with the “law of capitalism” that r>g, we have Pikettarians saying that, of course, Piketty never said that r>g explains the rise in inequality in recent years that concerns everyone, and in fact anyone who thinks he said that is a knave who has not read the book. I was among the many who never made it through Capital (it gave a new and different meaning to the expression “widely unread”), so I will take it on faith that the whole r>g thing was a head fake. Anyway, Capital does not make my list.

I think that number 1 is Complexity, by David Colander and Roland Kupers. On many pages, I highlighted insightful passages. On many other pages, I highlighted irksome passages. Look for a longer review from me next year.

Probably number 2 is Trillion Dollar Economists, by Robert Litan. It is a great achievement, but even so I wanted a different book.

Number 3 might be Isabel Sawhill’s Generation Unbound, about the pathology of unwed motherhood (it’s not just for teenagers any more) and what to do about it. However, I think that the question of whether “society” should be trying to prevent births of a certain type (namely, from unplanned pregnancies) is more difficult than she makes it out to be. Again, I have a review forthcoming.

Number 4 might be Charles Calomiris and Stephen Haber, Fragile By Design. The financial crisis continues to stimulate books on banking and related topics, and of the recent lot I thought this one had the strongest historical and international perspective. However, in the end, I found its main thesis, that U.S. banking policy is hindered by populism, unpersuasive.

Number 5 might be Mark Robert Rank, PhD, Thomas A. Hirschl, PhD, and Kirk A. Foster, PhD, Chasing the American Dream, which provides a good empirical study of income dynamics using longitudinal data.

Finally, there is a category of books written by friends of mine, in which I recommend Russ Roberts’ How Adam Smith Can Change Your Life, Megan McArdle’s The Upside of Down, and Elizabeth Green’s How to Build a Better Teacher.

UPDATE: Here is Tyler’s list.