Sunday, December 9, 2012

Weekend Links

  • The EIA projections of future US carbon emissions keep getting lower (graph above).
  • Bedpans and canasta in our future?  I worry about exactly this too.
  • Italy's prime minister Monti to resign.
  • Interesting paper on past ice sheet retreat (h/t Chris Reynolds). However, the money quote remains: "The rate of radiative forcing also needs to be considered with these palaeo ice-sheet analogues. The increase in boreal summer insolation and rise of 90–100 ppm in atmospheric CO2 during terminations occurred over ~11,000 years. In contrast, human carbon emissions have increased atmospheric CO2 by over 100 ppm in less than 200 years, with even greater increases predicted to occur in the coming century. Given this more rapid rise in greenhouse gas radiative forcing, Earth’s remaining ice sheets could respond in a manner not previously observed in the late Quaternary."


Chris Reynolds said...


Stewart, well spotted, it's been some time since I read that paper and I'd forgotten that caveat.

Stephen B. said...

Kevin Drum's piece is *exactly* the kind of thinking I've been kicking around for the past year or so only I don't see this path of increasing automation ending anywhere good via Drum's revolution or whatever eventually goes down.

I too, up until recently, was really against government entitlements being doled out too plentifully as I thought that the increasing lack of employment was simply the result of government initiatives making people lazy. Now, however, I see the decreasing employment rate as the outcome of automation. (Automation and technology can manifest themselves in many ways. It isn't just robots making things, but products like composite or vinyl siding lasting longer and putting wood siding people out of work or legal software, allowing 2 lawyers to do the work of 10. The list is nearly endless as to how changing technology is obsoleting jobs.)

I am reluctantly coming to favor more wealth redistribution programs, as the alternative as I see it, is automation, capital, and technology making a very few people very wealthy with the rest out of work. It will never happen completely as there are some service jobs that are hard to automate, but we are getting ever closer.

Automation is indeed giving nearly all of us more free time, only for a great many people, it is not by choice and if it hasn't affected most readers and writers on this blog yet, well, time will tell.

Like Drum, I have no real solutions as wealth redistribution gets very messy. Neither, however, is some worldwide Luddite movement likely if for no other reason than it would take an amazing bureaucracy to constantly police against any and all technological, job-destroying improvements. It would be something right out of a Orwell or Bradbury novel, though in many ways, we, of course, are already at a stranger place than those guys' fiction.

Stuart Staniford said...


You can see my thoughts at

I just reread it and I think it's holding up well so far.

Stuart Staniford said...

Krugman is starting to get on the case also:

Stephen B. said...

Stuart, I had missed that old post of yours and thank you for referencing it.

I think that your thoughts laid out there are a pretty good super set of what I've been kicking around. That said, I am not sure, as was discussed in the comments at the time, that all the problems of peak oil, climate change, and automation are all so neatly hierarchical, but I do agree with the general order.

I tend not to think of the Singularity as a singularity because I think things are going to break down in some way, before the majority of humans allow themselves to be relegated to the sidelines while the Elite's automata take over the economy, but then too, you said more or less the same thing on that as well.

Good analysis!

Seth said...

I was excited to read Krugman this morning too. You are the early-early warning, while Dr K is the earlyish warning system for everybody who reads the NYT. Then there's everybody else:(

Mike Aucott said...

It seems unlikely that robots, automata, etc. will interface seamlessly as they evolve as if through some grand design. Why would they not be subject to the same pressures of evolution as other organisms and systems? Once they achieve a position of relative independence from humans, and have the ability to control energy and other resource flows, they'll get to fight it out, and/or evolve mutually interdependent relationships with others of their ilk, just like others species do and have done. There will probably be many fits and starts in their evolutionary path. Who knows, robots may kill each other off and go extinct before humans do.

sunbeam said...

I don't have the slightest idea what an artificial intelligence would be like, and I kind of think no human being can meaningfully speculate either.

Take as basic a concept as self preservation. Is it a guarantee that an artificial intelligence that was created without the instincts we have, care if it were threatened with destruction?

"My off switch? It's right over there. Hmmm you are going to pull it? Oh well, I'm going to keep running these calculations until I am off."

No AI that will be built will feel any motivation from sex. What percentage of human activity is due to this motivation alone?

Is it possible to build an AI that is not self aware? What does that mean exactly anyway?

Just saying that any intelligence we might produce is inherently unknowable to us at the current time. Maybe unknowable in any conceivable sense as well.

I've read lots of science fiction and seen movies about Skynet. My thinking is that a true AI wouldn't necessarily "care" about us even to want to destroy us.

My immediate trepidations about this whole field are pretty common to everyone that posts here however. You don't really need to produce some science fiction AI supercomputer that is smarter than we are to change everything.

Greg said...

RE: Bedpans and canasta

The "robot revolution" is not just replacing people one-for-one on the job, but also replacing them with household appliances. Perhaps the revolution started with the vacuum cleaner...

This may arrive sooner than the self-driving car:-

[SCOUT] is a tiny hardware device that reads your vital health information on contact. You simply place it on the left temple and, in less than ten seconds, it will read your pulse transit time, heart rate, electrical heart activity, temperature, heart rate variability and blood oxygenation. Then it sends this information to an app on your iPhone or Android phone, which displays it for you. You can even store your vitals for tracking, which could prove fundamental to many health situations at home... will cost around $150 when it appears at the end of 2013, after it gets US government approval.


ScanaFlo,..a disposable blue plastic rectangle with a QR code...[when it] get[s] a reading...the swatches will change color.

But what do these color [sic] mean? You don't have to guess or remember. Point your smartphone at the QR target and it will...tell[..] you if it detects anything out of the ordinary. According its creators, ScanaFlo tests for "pregnancy complications, preeclampsia, gestational diabetes, kidney failure and urinary tract infections."

ScanaFlu works in a similar way. ...Incredibly enough, this "disposable cartridge will provide early detection for Strep A, Influenza A, Influenza B, Adenovirus and RSV." Like ScanaFlo, you will use your phone's camera to have a result sent to your app.

These disposable systems will be sold in packs, also at the end of 2013.


If these gadgets...[work]...[t]he monetary savings in prevention alone—and not depending on expensive laboratories for many tests—makes it all worthy.

But even more exciting is the potential increasing accuracy of diagnostics, based on the tracking of data over time...

I think it likely that there will be demand for these devices in the USA.

Near-term descendants of these devices, and similar ones such as the eye-testing smartphone app-and-attachment developed by some MIT students, together with associated dispensing systems, will greatly reduce the required number of doctors, nurses and technicians around the world, while improving medical practice.

Innovations like these will be a great boon in developing countries such as Tanzania (which is predicted to be the fourth biggest country in the world by 2100, but still poorer than China is today).

But they will also reduce the number of possible ways to make a living.

Stuart Staniford said...

At least for the forseeable future, computer intelligence will be delivered by software companies which will optimize it and carefully test it to solve some problem which their paying customers want solved. So the goals of AI will be profoundly shaped by the goals of capitalism, and in particular its financial elites. I don't believe AI's will be running amok any time soon, they'll just be slowly displacing more and more people from the economy (which is the explicit goal - in the biz we call it ROI).

Sam said...

How would a future, where mass manufacuring is automated and thus there are no jobs for people, actually work economically? Who is going to be buying all that stuff??? Not the people without jobs, and even if you're super rich you can only buy so many things...instead you typically buy a few very expensive things that are not massed produced. Our economy is based on selling mass goods to the masses, if there are no masses, there is no money to made and then all that expedeniture in capital to automate mass production is a waste meaning the rich people lose the value of their investment and become poor too. So I'm not sure where the incentive is that would lead us to a world where everything is automated...

of course I see things like the maker movement, the local movement, tesla, the first successful private ventures into space, climate change...I can't predict what the future will look like but it seems like there is going to be lots of new opportunnities and that being replaced by robots with nothing left to do but twiddle ones thumbs is not a likely one.

Manolo said...

Regarding the Anders E. Carlson
and Kelsey Winsor paper.
The catch phrase that should be rising all red flags :"Given this more rapid rise in greenhouse gas radiative forcing, Earth’s remaining ice sheets could respond in a manner not previously observed in the late Quaternary."
This seems very obvious, and most researchers underestimate the system sensitivity of the cryosphere by a very wide margin, IMVHO. This is going to bite us.
The CH4 (Methane) problem is developing at an VERY alarming speed,sort of step linked to the sea ice volume lost in the Arctic, confirming all our fears, see link:>
Well, if this is not a "RATIONAL ANALYSIS OF GLOBAL CIVILIZATIONAL RISK" i do not know what is...

Stuart Staniford said...

Sam asks "How would a future, where mass manufacuring is automated and thus there are no jobs for people, actually work economically? Who is going to be buying all that stuff??? Not the people without jobs, and even if you're super rich you can only buy so many things"

I think you can make a case that we are already dealing with the beginnings of this - we have labor income a falling share of GDP, we have median income flat/falling for a few decades, and then we have consumers having overborrowed to maintain their lifestyle, culminating in the housing bubble. Now we have a very low interest rate environment (corresponding to the phase where the rich can't make a decent return on their investment because their isn't enough mass demand for new products).