Did the Singularity Already Happen?

Ezra Klein writes:

Ezra Klein - Our depressing robot overlords: Paul Krugman writes, “the idea that modern technology eliminates only menial jobs, that well-educated workers are clear winners, may dominate popular discussion, but it’s actually decades out of date.”... [T]he difference between what computers can do and what computers cannot do is not whether the job requires a college education, but whether doing the job can be broken into “routine and repetitive” tasks. Martin Ford, who has done some thinking on these issues, draws out the implications:

The key thing to understand here is that our definition of what constitutes a “routine and repetitive” job is changing.... As specialized artificial intelligence applications... get better, “routine and repetitive” may come to mean essentially anything that can be broken down into either intellectual or manual tasks that tend to get repeated.... [I]f 50% of a worker’s tasks can be automated, then employment in that area can fall by half. When you begin to think in these terms, it becomes fairly difficult to make a list of jobs that (1) employ large numbers of people and (2) are completely safe from automation.

The obvious set of questions this raises is “how will the economy adapt?” Krugman argues that “if we want a society of broadly shared prosperity, education isn’t the answer -- we’ll have to go about building that society directly,” perhaps through things like unions and universal health care....

But I’d pose a different question: How will we adapt psychologically?... How do you keep morale up in an economy when more people are simply less necessary than they used to be?

That’s a harder question to answer than “how do you make sure everyone has access to medical care?” But for a substantial fraction of the population -- not a majority, but certainly millions and millions of people -- it’s an increasingly pressing one. People get trained for a job in their 20s, and then, in their 40s, that industry gets disrupted by technology, or sent to China, and even if some of those people find jobs again, they tend to be at a lower level -- a drop in status and perceived usefulness that’s psychologically devastating. This is a question for not only the future, but given the number of long-term unemployed in the economy right now, the present. And it’s not a question that we have any very good answers for.

We have gone from having 80% of our males working outside-the-home as farmers and farm laborers down to 2%. We have gone from having a world in which nearly every female past puberty spends her life effectively chained to her children, her kitchen, or her loom to one in which I cannot think of a house I have been in in a decade that had a loom. We have already gone through the great transformation by which the general business of life--growing and processing our food, building our shelter, weaving our clothes, and telling ourselves stories for information and entertainment--has been extroardinarily, comprehensively automated. And yet we have found things to do.

Consider the room that I am now in as I wait for my 12:00 noon PBM down here in Durant Hall. In half an hour there will be twenty people in the room. There will also be lunch for 20 (market value $200), 10 tables (market value $1000), 30 chairs (market value $1500), one projector (market value $1500), one carpet (market value $500), and one conference room (market value $200,000). Given amortization rates of 3 years for the projector, 25 years for the building, and 5 years for the tables, chairs, and carpet, and assuming the conference room is used 4 hours a day, the economic activity in the room for our hour-long meting will consist of:

• $14.00 for the room
• $1.00 for the tables, chairs, and carpet
• $0.50 for the projector
• $200 for the lunch--of which $40 is the farm gate price of the ingredients, $100 is food processing and preparation, and $60 is delivery to the right room at the right time.

Figure that the people making these things are earning an average of $10/hour, and we have 22 hours' worth of work that will be contributed to the process. We will add another 20 hours as we discuss Berkeley's Gen Ed curriculum. A normal economist would say that you should value our time at the $50/hour that the market values it: since the meeting ran over to an hour and a half, and counting our preparation time for the meeting, that is $3000 of academics' time.

Total economic value added produced in the Durant Hall Basement Conference Room at Monday lunchtime was $3215.50.

Of this, 1.7% was food and shelter. 5.0% was making the food tasty and delivering it on target on time. 0.05% was for items--tables, chairs, carpet, projector--to make our meeting more productive. And 93.3% was the work of attempting to plan and revise U.C. Berkeley's breadth requirements.

Contrast this with a similar academic meeting taking place at Sorbonne 900 years ago with the same purpose--to revise the then Gen Ed program: logic, grammar, rhetoric, music, arithmetic, geometry, and astronomy. Farming today is 30 times as productive as it was back then. Construction today is at least 20 times as productive. Furniture and cooking are perhaps 10 times as productive. Transport and delivery within the university? Call that 1:1. I don't want to think of the cost of an overhead projector equivalent--you make do with parchment, goose quills, and inkstones. Say 200 times as productive? And we academics? Pick a number between 1 and 5 for how we are able to apply what we know to produce knowledge--I don't have a guess.

Assert the same 5:1 ratio of academic wages to average wages. It probably wasn't much more than that: medieval academics were poor. Remember Chaucer's Clerk of Oxenford: "[h]e nas nat right fat, I undertake/But looked holwe, and therto sobrely./Ful thredbare was his overeste courtepy;.../For hym was levere have at his beddes heed/Twenty bookes, clad in blak or reed,/Of aristotle and his philosophie,/Than robes riche, or fithele, or gay sautrie./But al be that he was a philosophre,/Yet hadde he but litel gold in cofre;/But al that he myghte of his freendes hente,/On bookes and on lernynge he it spente"? Remember Desiderius Erasmus: "When I get a little money I buy books; and if any is left I buy food and clothes"? They weren't kidding: "In Byzantium... around 900...Arethas, Archbishop of Caesarea value[d] his copy of Plato at 21 nomismata... and Euclid at 14 nomismata (perhaps not including the parchment). Manual workers in Byzantium were paid 6 to 10 nomismata per year."

Then for the Sorbonne 1110 Breadth Requirements meeting we have:

• Room: $280
• Furniture: $10
• Parchment, etc.: $100
• Calories: $1200
• Food processing: $1000
• Delivery: $60
• Academic labor: $3000

A total of $5650 is the cost of having the same lunch meeting. Academic labor is 53%. Bare calories and shelter are 26%. Food processing and delivery are 19%. And that leaves just a smidgeon--under 2% of the meeting cost--for furniture, parchment, inkstones, etc.

Even in an institution as rarified as one of the world's leading universities, the share of economic activity of even the most-high powered meeting--multiple deans, Nobel Prize winners etc.--that was simple basic food and shelter was about 50% a millennium ago, and is down to 2% today.

In short, the automation revolution already happened. And we adjusted.