Congratulations to the people of the SIAI for putting on an interesting, stimulating, really well-run conference. Note for those who weren’t able to be there, that the talks were recorded and will be available on SIAI’s website.
The following comments are not an attempt to explain the talks but simply to record a few odds and ends of thought that occured to me while listening to them.
In Brad Templeton’s talk, a fascinating account of the possibilities of self-driving cars (“robocars”), he lists as one of the drawbacks of having more efficient cars the possibility that people might therefore take longer trips. They will of course do this; it’s probably the simplest and most robust prediction of basic price theory that when the price of something falls, people will demand more of it. The really interesting possibility is something called the Jevons paradox: as energy (or whatever — Jevons wrote in the 1860s about coal use) becomes efficient, people use so much more of it that the total amount used increases; hence why it’s called a paradox.
It’s not really a paradox in any sense except that the average person’s intuition doesn’t exend to price theory. The determining factor in whether the total energy use will rise is whether the demand for energy use is inelastic (steep demand curve) or elastic (shallow demand curve). Another way of looking at elastic demand is that the “before” price is just barely too high to afford something valuable; when it drops just a little bit, a lot more of the valuable thing will be bought.
Now the reason I think that Brad’s robocars will involve a huge Jevons effect is that a robocar is not just a little less expensive than a current-day car — it’s enormously less expensive. It’s less expensive in energy, by some amount; but it’s also less expensive in lives, and its hugely less expensive in time. Brad estimates the time savings to be worth on the order of a trillion dollars a year.
But the value of getting from point A to point B will be just as great. Now I know from my own experience that my travelling is hugely elastic to cost, particularly time cost. So I expect a major Jevons effect from self-driving cars.
Some people talk about the Jevons paradox, especially in energy, as if it were a bad thing: we went to all that work to increase energy efficiency and you fools went and increased total energy use, you horrible greedy people you. But it’s in reality just the opposite. The reason that the new usage happens is that it’s valuable. When the Jevons effect operates, it means that not only has money (time, resources, lives) been saved on the old uses, but that new uses are made possible that have a value to society greater than the savings!
Thus if robocars would save us a trillion dollars of wasted time with the current amount of driving, they are likely to enable more than a trillion dollars of totally new transportation. And that’s a stimulus that would actually work.
More on other talks in future posts…