Saturday, December 29, 2012

Saturday Links

  • Paul Krugman has been getting much more interesting lately looking at the economic signature of the rise of smart machines.  My take from a couple of years back is in the back half of this essay. At the time I wrote that, I felt like I was the only person with my views (possibly wrongly, but that's what it felt like), but lately Kevin Drum and Paul Krugman have both started to espouse somewhat similar positions. 
  • The anatomy of US partisanship.  My understanding of the political science literature on this is that it's more due to people self-selecting to live in ideologically compatible places than it is due to gerrymandering (though there's an assist from gerrymandering).
  • US civil rights continue to suck.
  • Outsourcing is hurting Japanese cities and workers too.
  • This NYT article is sort of interesting and sort of infuriating.  It concerns how utilities in the northeastern US are thinking maybe they should spend a lot more to harden the electrical grid in light of the storms of the last two years.  It's clear the entire issue is driven by climate change, but the article manages to avoid ever mentioning that, instead trading in vague circumlocutions like "Utilities and policy makers can see that ocean surge poses a previously unexpected threat to the power grid." and "forecasters say that so-called 100-year storms are likely to occur more frequently".  Gutless weasel words: the article should make clear that we may have to spend a lot more to make the electric grid robust and this is a cost of human-caused climate change.
  • And here is the EIA talking about possible exploitation of polar oil and gas resources again without any mention of why this is now perhaps becoming feasible. 
  • Interesting survey post on the impacts of oil prices on the economy.  
  • On a side note, that last link is a bit daring right now, since author Michael Levi is at the Council on Foreign Relations and their website just had a major compromise by foreign intelligence   resulting in multiple compromises of US companies (what we in the security industry are calling a waterholing attack).  I'm aware of this because software I wrote for my day job detected the compromise :-)  Anyway, their website is now cleaned up (and will hopefully stay that way), but if I had inadvertently linked you there last week, you might have gotten infected.  I'm aware of at least six distinct waterholing compromises of US think tank websites in the last three months.


Robert Wilson said...

I first encountered Norbert Wiener's book Cybertetics while reading Time Magazine in 1948. His second book was:

Lucas Durand said...

Re exploitation of arctic fossil fuel resources - Tad Patzek has made an interesting post on his blog:

James said...

Congradulations on your software detecting the compromise!

James said...

I think it is worth pointing out that Krugman is concerned about the effect of technology generally (including the internet), and not just the rise of Smart Machines.

He discusses his thinking around the "superstar effect" in detail here:

Robert Wilson said...

I should have said more about Wiener, father of Cybernetics. He discussed the potential benefits and possible harm from automation 60+ years ago.
From Wiki:
"...argues for the benefits of automation to society. It analyzes the meaning of productive communication and discusses ways for humans and machines to cooperate, with the potential to amplify human power and release people from the repetitive drudgery of manual labor, in favor of more creative pursuits in knowledge work and the arts. He explores how such changes might harm society through dehumanization or subordination of our species, and offers suggestions on how to avoid such risks."

Mike Aucott said...

Krugman, in his 12/28 piece in the NYT, “Is Growth Over” does seem to be coming to the conclusion that it will take more than the world’s central banks printing more money to bring back a comfortable economic situation. But, he still does not seem to get the idea that energy constraints, especially the lack of cheap liquid fuel, are a significant, probably critical, part of the current problem. Krugman mentions Robert Gordon’s recent article (available at ) without noting that Gordon’s entire discussion mentions energy only in the context of its likely future restrictions due to the need to limit carbon emissions. Gordon completely misses the point that much, if not virtually all, of the dramatic economic growth over the last three centuries coincides closely with, and is almost certainly inextricably linked with, the exploitation of the vast energy of fossil fuels. If the era of cheap fuel is coming to an end, it’s a whole new ball game economically. This is a point that Krugman, and indeed most economists, seem to consistently fail to grasp.

It may be true that smart machines will help us do more with less and so maintain economic growth. But, given a fairly certain positive correlation between growth in GDP and growth in energy supply, and the virtually certain correlation between carbon emissions and potentially catastrophic climate change, these machines will have to be smart enough to help us grow the energy supply while at the same time drastically cutting carbon emissions. Either that, or humanity will have to get serious about developing a steady-state economy, which will almost certainly require something that we humans aren’t always good at - sharing.

sunbeam said...

Mike Aucott said:

"It may be true that smart machines will help us do more with less and so maintain economic growth."

I think that economic growth for a select sector of the population will be the only thing of interest to the powers that be.

Even besides that being a planned objective, can you see any other way things will play out, given the parameters of what we are working with?

Stuart kind of ... pooh pooh's might be the wrong expression, but seems to think that that a lot of people are too extreme in worrying about the effects of robotics, or at least the time frame.

I tend to think things can happen very quickly. And the end result is that we are at the cusp of a time in which labor has no value. At least in the aggregate. Obviously someone will have to create new applications and implement things (let's not deal with extreme "singularity" issues), but I think we have more than enough skilled people to do that now. And even if it were a bottleneck, the potential savings dwarf any premium you would have to pay to individual programmers.

So what then? Look I've been thinking about things recently. All it takes is for someone in the world to solve a specific problem, say a truly effective robotic grass cutting system.

Potentially the same system could be implemented everywhere in the world. Truthfully the most important thing needed would be the proof it is possible, and a general idea of how they dealt with the problems, in order to make something similar.

And that is without the knowhow of such a thing being freely shared.

sunbeam said...

I thought of something else I wanted to add.

What if we are in a economic situation that is unlike those in the past? That's pretty obvious I guess, but what concerns me is Keynesian pump priming.

Let's say that the economy has multiple "stable" states, and that the purpose of Keynesian spending was to kick it into a state you preferred, over another equally stable state.

What if the feedback mechanisms have changed such that you can't kick it into what you want anymore?

What I'm trying to say is, what if it doesn't work anymore? You have stimulus spending, things move the way you want for a little while, then they move back to the state you don't want, because that is what the underlying mechanisms favor.

I'd have to do a little reading to couch this argument in economic terms, but has anyone ever looked at this?

There are other economic models (communism, feudalism, etc), but if we can't use Keynesian spending to get where we want to go, what exactly does that mean? The US is pretty much married to it's current system, more so than it's married to Christianity or any other belief system. I can't see it being able to change too much.

Robert Wilson said...

My favorite example of a job stolen by a smart machine was that of elevator operator. "floor please - step to the back" Other obsolete jobs include
The old fashioned country club caddy lost to a dumb machine - the golf cart.

Greg said...

I've just come across Jorgen Randers's new book, 2052 - A Global Forecast for the Next Forty Years. The website for the book is at There's a 5000-word PDF summary there, and a better summary with charts at Cambridge Programme for Sustainability Leadership (scroll down, or search on the page for "Randers"). Randers has made available a spreadsheet (Excel 2007 or later) with his numeric data and charts at

Randers, you recall, is one of the authors of Limits to Growth. This new book is explicitly a forecast (a single forecast) rather than a set of scenarios like LtG.

In some ways the forecast is quite optimistic. Randers sees fossil energy use and CO2 emissions peaking at about 2030, and food production easily keeping up with population growth and increased affluence. Water scarcity will be a smaller problem than some think, too. Due to increased investment, unemployment won't be too bad.

In other ways, 2052 is more pessimistic. Randers expects that wilderness will be essentially gone by then; we'll experience "wilderness" in carefully managed parks, which will shrink as climatic zones move polewards. Economic growth will be lower than conventional estimates, and we'll have to use an increasing fraction of GDP for resource extraction and to combat the effects of climate change and pollution. (This is the investment mentioned above.) A bigger fraction spent on these things leaves a smaller part for household goods and services. So in 'advanced' countries, growth will feel like stagnation at best.

And Randers thinks that we won't have solved the big problems of topsoil loss and climate change. While he doesn't discuss it, he thinks the period after 2052 will be worse than the next 40 years.

The basic problem, as Randers sees it, is the short-termism built into both capitalism and democracy. We won't do anything significant about problems until they're already on us, and even then may do the wrong thing. (Given the current drama about US Federal revenues, you'd have to say he's got a point.) Other important assumptions are that fertility will continue to decline; we won't develop effective supra-national bodies (a UN that works); and that the 120-odd "less developed countries" won't solve their governance problems: elites in those countries will continue to act as robber barons towards their own people, or lose control entirely.

There are omissions, at least from the summaries. Accelerated automation and its potential effect on incomes are not mentioned, for example.

Anyway, Stuart, this might be worth casting your eye over. I, for one, would appreciate your comments.

Nick G said...


You might want to look at the work done in conventional economics on the problems of mature industries, obsolete professions, and underemployment of those with obsolete skills.

I believe that these are pretty old problems, and that they've been a central topic of economics ever since skilled craftsmen were displaced by the AI of the 1700s and 1800s, such as automated looms.