Tyler Cowen writes a book about how he was wrong
Apparently I never wrote a review of Tyler Cowen's The Great Stagnation, which is weird bc Marginal Revolution is one of the most intellectual blogs I read and I found the book very novel. The premise is that America has eaten the low hanging fruit of cheap land and labor force expansion, so future economic growth will have to come from scientific developments, which aren't turning out to be as profitable as you would think. Two MIT economists, Erik Brynjolfsson and Andrew McAfee, disagreed and wrote a short rebuttal called Race Against the Machine. I kept meaning to read the latter, but didn't, which is okay bc Cowen must've.
Average is Over is Cowen's mea culpa, in which he suggests that machine learning will make everything better. Brynjolfsson and McAfee call this the second digital age, in a book of the same name that I have also failed to read yet, even though Sloan gave me a hardcover copy a few months ago. Just as the Industrial Revolution increased our physical abilities with tools that multiplied our force, the digital revolution increases our mental abilities with tools that multiply our faculties.
Cowen's biggest example of this is freestyle chess, in which players have a set amount of time, but can use ANYTHING they want to come up with the next move, up to and including multiple computer chess programs. Cowen calls chess the Drosophila of intellectualism and the results are... interesting. The winners of these competitions are decent chess players -- who are really good with computers and can interpret which program is most likely to be right in a given situation.
This is one of the most important points for Cowen, Brynjolfsson, and McAfee: just as John Henry was much less important in the age of steam, being "smart" is much less important in a world where computers can think better than you. Cowen talks about how computer chess programs come up with really weird looking moves that work anyhow: as more tasks get machine-outsourceable, there'll have to be a lot more faith in the weird leaps that machine learning makes. Brynjolfsson & McAfee talk about how as late as 2004, they were still teaching at MIT that fine motor skills, language processing, and pattern matching were too hard for computers. HA! What other cognitive problems will they solve in the next ten years?
I had a hard time believing Cowen’s explanation that machine learning will eventually replace scientific papers. I do think he’s right that computers are going to continue climbing the curve of human “thinking”, with a gradient that is determined by how much of the task requires considering the mind state of other humans.
So far this is good for me and my tribe: if you're on LiveJournal, you've been using computers for a long time. The bad news is, being less smart but more conscientious, driven, and having more self-control will be preferable in this brave new world. It doesn't take all that much intelligence or training to personalize your iPhone, use GPS, or consult Google more than doctors, but managing a team of people managing a bunch of computers will be hard enough that employers won’t want to waste time & energy on a cheap employee unless they are extremely reliable. Moreover, education will be cheap enough that anyone who is slightly motivated can learn as much as they want over the Internet. I think it was Tyler Cowen who first commented in my "hearing" that the Internet really bifurcates those who are driven - because they can learn anything and everything over it - versus the pikers who use it for Facebook or playing games. Ahem.
The real problem with all of this is that, just like the Industrial Revolution, the returns will go primarily to capital holders: there’s a huge swath of the developed world that will have a harder and harder time getting the sort of work they are used to doing, for the pay they expect. Europe has already had its first taste of this, and it hasn’t been pretty. Brynjolfsson & McAfee actually did a lot better job of explaining this at the last talk I went to, but essentially more and more of what is valuable has a marginal cost of zero: the millionth copy of Flappy Bird costs no more than the first, so if you are motivated to make a game that goes viral, you take all. Cowen argues that we will become more like Mexico: basic life will be cheaper and we’ll be happier because entertainment is cheap, but the haves will have a LOT more.
Someone once commented about my reviews that he didn’t need to read the book after reading my review. This is one case where the book contains far more than I could reasonably summarize: I couldn’t even get into his comments on healthcare and education in the new world. If you’re interested in where the world will be in 20 years, this is a book well worth reading.
Average is Over is Cowen's mea culpa, in which he suggests that machine learning will make everything better. Brynjolfsson and McAfee call this the second digital age, in a book of the same name that I have also failed to read yet, even though Sloan gave me a hardcover copy a few months ago. Just as the Industrial Revolution increased our physical abilities with tools that multiplied our force, the digital revolution increases our mental abilities with tools that multiply our faculties.
Cowen's biggest example of this is freestyle chess, in which players have a set amount of time, but can use ANYTHING they want to come up with the next move, up to and including multiple computer chess programs. Cowen calls chess the Drosophila of intellectualism and the results are... interesting. The winners of these competitions are decent chess players -- who are really good with computers and can interpret which program is most likely to be right in a given situation.
This is one of the most important points for Cowen, Brynjolfsson, and McAfee: just as John Henry was much less important in the age of steam, being "smart" is much less important in a world where computers can think better than you. Cowen talks about how computer chess programs come up with really weird looking moves that work anyhow: as more tasks get machine-outsourceable, there'll have to be a lot more faith in the weird leaps that machine learning makes. Brynjolfsson & McAfee talk about how as late as 2004, they were still teaching at MIT that fine motor skills, language processing, and pattern matching were too hard for computers. HA! What other cognitive problems will they solve in the next ten years?
I had a hard time believing Cowen’s explanation that machine learning will eventually replace scientific papers. I do think he’s right that computers are going to continue climbing the curve of human “thinking”, with a gradient that is determined by how much of the task requires considering the mind state of other humans.
So far this is good for me and my tribe: if you're on LiveJournal, you've been using computers for a long time. The bad news is, being less smart but more conscientious, driven, and having more self-control will be preferable in this brave new world. It doesn't take all that much intelligence or training to personalize your iPhone, use GPS, or consult Google more than doctors, but managing a team of people managing a bunch of computers will be hard enough that employers won’t want to waste time & energy on a cheap employee unless they are extremely reliable. Moreover, education will be cheap enough that anyone who is slightly motivated can learn as much as they want over the Internet. I think it was Tyler Cowen who first commented in my "hearing" that the Internet really bifurcates those who are driven - because they can learn anything and everything over it - versus the pikers who use it for Facebook or playing games. Ahem.
The real problem with all of this is that, just like the Industrial Revolution, the returns will go primarily to capital holders: there’s a huge swath of the developed world that will have a harder and harder time getting the sort of work they are used to doing, for the pay they expect. Europe has already had its first taste of this, and it hasn’t been pretty. Brynjolfsson & McAfee actually did a lot better job of explaining this at the last talk I went to, but essentially more and more of what is valuable has a marginal cost of zero: the millionth copy of Flappy Bird costs no more than the first, so if you are motivated to make a game that goes viral, you take all. Cowen argues that we will become more like Mexico: basic life will be cheaper and we’ll be happier because entertainment is cheap, but the haves will have a LOT more.
Someone once commented about my reviews that he didn’t need to read the book after reading my review. This is one case where the book contains far more than I could reasonably summarize: I couldn’t even get into his comments on healthcare and education in the new world. If you’re interested in where the world will be in 20 years, this is a book well worth reading.