Lance Knobel: So, Brad, one of your areas is economic history. I am curious: as we face this increasing automation, robotization, is this something that’s likely to be something we have seen before in economic history or is this time going to be different?
Brad DeLong: Well, it is always going to be different, because history does not repeat itself--although it does rhyme. The question is: how is it going to be different?
Looking back at all the major transformations in history before--as we have seen entire categories of things we do to add value to our society vanish--we always found new valued things for people to do. Technological unemployment has been a yearly thing, a decade thing, a generational thing perhaps--but never before more than a generational thing.
At times, the consequences of change have been horrible for the standard of living of the average guy. Jared Diamond has a nice line about how the invention of agriculture was the greatest mistake in the history of the human race. Iit allowed us to support 100 times more people on the planet than we could by hunting and gathering, yes. But the average life of an agriculturalist, staring at the hind end of an ox for eight hours a day--if they are lucky, and pulling the plough themselves if they are not--is not a very rewarding use of the incredibly flexible cognitive instrument that is the human brain.Yet that’s what most people wound up doing for most of the time between 5000 BC and 1700 or so.
Moreover, there is the fact that your average hunter-gatherer back in the Neolithic was 5’8" or so for adult males, and your average agricultural peasant in China or in Europe or even in America was 5’2". I don’t know about you, but if I fed my children a diet that would make them 5’2 at adulthood, Alameda County Child Protective Services would come and take my children away and I would never see them again.
The fear that this time may be very different arises because in the past the human brain has been a uniquely effective control and guidance mechanism for all kinds of things. The ear, eye, brain, hand, voice loop is something hitherto unmatched. Plus: the human smile and touch and glance is an extraordinarily effective social mobilization device for getting people to pull more or less in the same direction.
We are looking forward into the future in which people are going to find ways to make machines do huge amounts of tasks that have been reserved for the human ear-eye-brain-hand-voice loop in the past. I don’t know about your children, but I look at mine watching anime with the big googly eyes, and I wonder how long the human smile and glance is going to have its favored place as well.
Lance Knobel: So Josh, [inaudible]. And, even more pertinent, what is already here that’s going to kind of increasingly become part of our lives?
Josh Bloom: When I think about robots, first of all I should be clear that I am not thinking about sort the mechanical robot, the mechanization and the automation of factories. Of course that’s hugely important. One could argue that the existence of that process, while displacing ordinary factory workers, has been making the ones that remain more efficient, making them more valuable and more productive and then creating vast amounts of wealth in other places of the economy.
When I think about robotization, I am thinking about the sort of cognitive automation, doing just what Brad was mentioning is now possible. We can now build machines that can do a lot of the classifying and learning and being able to make predictions about the future that have a real impact in our lives.
We are already started to see these software bots make a difference in our own lives. I am sure all of you have become more comfortable opening up your inbox because there is less spam now. If not, you should probably switch email providers. Spam is essentially a finished product, doing spam protection is a well-understood task, it’s a machine looking at your mail before you even see it and making a decision of is it spam or not, and it putting it one place if it is and another place if it’s not.
Credit card transactions are now looked at with artificial intelligence. We are getting to the place now where cancer detection is becoming automated via bots looking at cells. And not just doing what humans could have done themselves--only doing it faster--but now doing it better. When we start thinking about teaching machines how to replace us, there is a huge negative that immediately comes up.
But if we think about teaching machines how to inform us as a way of helping us and becoming personal assistants to us--way beyond what we think about now, becoming a trusted confidant, it’s very clear that this type of work is going to have a major impact. What I think is useful point out is not just that this is in our own [inaudible], but it’s only getting better.
It’s something that we can talk about. But it is amazing if you think about it. Where we are right now, here in Berkeley, on our campus just a few blocks away--some of the major technological advancements were invented there that are allowing us to understand the world, and allowing us to start thinking about creating this automation of knowledge...
Brad DeLong: When I listen to Josh Bloom I think: how rarified is the air we breathe here! What we regard as a task that ought to be too boring to force graduate students to do--that is, looking at pictures of pieces of the sky looking for something weird--is itself much more interesting and much more a fair and effective and proper use of the human cognitive super computer than 90% of what actual white collar workers do out there in the world today. 9-to-5.
It’s a much more interesting task than, say, deciding whether applicant X qualifies for Medicaid, or whether applicant Y has properly managed to fill in the names of the family's dependents in the right spaces. And then thee are the tasks that we have already seen vanish--40 years ago we had half a million women at telephone switchboards plugging cables into receptacles in order to make switches and connections. That was a huge part of what people did. We are going to manage to eliminate all of that, and most of that will be extremely good to be gone--as long as we can find ways of getting the people who would have done that their incomes, and finding other things to do so that they feel useful and wanted and productive, and not like parasitic members of the moocher class...
Brad DeLong: And there is the question of how much of this gets into our national accounts. One example: back in the 1980s I was incredibly pleased to have the opportunity to purchase an Encyclopedia Britannica for only $1000. We gave it away the last time we moved, because no one had looked at it in years. Wikipedia was simply so much easier and so much better. To think that I was willing to pay $1000 for an Encyclopedia Britannica, and now we have universal scale by every household to Wikipedia--scale that up and you’ve got $100 billion of extra wealth for the United States alone. That’s not quite the right calculation to do. But it is the same order of magnitude. And none of that $100 billion of wealth shows up in the National Income and Product Accounts anywhere...
Lance Knobel: But at the same time we have had this wealth [inaudible] the internet has created enormous wealth, created industries, seized all sorts of opportunities, the transition from Encyclopedia Britannica to Wikipedia. We have had increasing incoming quality and we have had a [inaudible] out of the middle class. So is this technologically driven, Brad, or are there other factors in this policy issue? Where do we stand on that [crosstalk] to continue that trend?
Brad DeLong: Well, at the moment I am going to take refuge with Emmanuel Saez six doors down the hall on the sixth floor of Evans Hall. I will say that as of now rising inequality has been a policy choice. Over the past generation we have seen rising inequality in English-speaking countries--in Anglo-Saxon countries. We have not seen it in non-English speaking North Atlantic economies. It is not even a Germanic language thing: it is the Anglo-Saxons who have seen greatly rising inequality over the past generations, while the Saxon-Saxons have not, and the Nordics have actually seen increasing equality over the past generation.
Male Speaker: And the Gauls?
Brad DeLong: And the Gauls are still doing fine with their 1000 kinds of cheese. It is remarkable.
But at some point at least the pressures making for rising inequality will become technological. Rather, technology will pose us problems of income distribution and social worth to solve. It’s just they haven’t shown up yet on a scale large enough to see in Emmanuel Saez's cross-era cross-country graphs...
Brad DeLong: Let me endorse that. Our current patent system was state-of-the-art, forward-looking social engineering back in 1787 when we put in the constitution. It was then a sign of how progressive we are. But by now our intellectual-property system is well past its sell-by date.
I think just made a decision. Let me make an announcement: Everyone taking Econ 2, Principles of Economics, next semester, listen up: if you take the class from me you are going to have to learn to do your problem sets in Python, so get ready...
Lance Knobel: I think there are so many different things that are potentially areas of questioning. But I want to ask one thing, Brad if you acquire Python for your economic students and you told Josh about to teach Python and it’s important for scientists and all that, do both of you have any [inaudible] or the rising generation of students, if they don’t code they are going to be excluded from a productive economic life? Is that going to be essential, or is this just for a technocratic slither of the economy?
Brad DeLong: Well, it would certainly give them some extra options that might not have otherwise. It’s actually unclear to me what people will be programming come 20 years from now. I tried to map what my daughter is learning in her introductory programming classes onto my brain--which I hate to confess is basically still a Fortran brain, that whenever I write something I say "how I can do this in Fortran?" And then when I show my code to the trained professionals, they cry. But we try. And certainly one thing a university exists to do is to give as many people as possible as many mental tools as possible. There are enormous amounts of mental tools, and they are useful in all kinds of places.
I know one person on Wall Street who says that the best thing for him in his entire education--after he learned enough math to be able to fake it while talking to his quant--was his history classes, with their focus on source criticism: why is X writing this in year Y to say Z? What purpose are they trying to achieve? And what does that tell us about how far we should trust them? He said that, for life in Manhattan these days, that’s the most important skill you can possibly learn...
Lance Knobel: We have a roaming mic. Why don’t we start right here.
Greg: Thank you. My name is Greg, I am 45 so I am old enough to be wired, but also able to disconnect myself--I think. Something that Brad touched on at the opening comments grabbed me. Certain things I am hearing now that they are finding that children as old as 10 or in that area, they are seeing that they are less able to interpret facial expressions and communicate in that way because of this world we live in where we are facing down looking at our smart phone, or whatever.
As that relates to not only human contact and emotional well being et cetera. There is even things about how when we make eye contact with people--even with our dog, that we are actually generating vitamin B [inaudible]. So as we move forward, this robotization of our world and interaction and the lack of need to exchange information and that’s both subconscious and conscious and all that, I wonder what your thoughts are about evolving. It’s a physical phenomenon, it’s an emotional phenomenon and it’s a cultural phenomenon.
Brad DeLong: Though I suppose one way to approach that question is to ask: Is the singularity in our future or in our past? And: Is it scary? Our lives are already so different from those of our hunter-gatherer environment of evolutionary adaptation. I am reminded of Isaac Asimiov's novels--The Caves of Steel and the Naked Sun--about his robot detective and his human partner, trying to deal with societies that had become so wired and so isolated that individual members could barely stand to be in the same room with each other, and found it completely gross and disturbing to actually have another human body within 10 feet of them.
These are books by a New Yorker of the 1950s, who feels that he is a bit too focused on technology and chemistry and atoms and so forth, and trying to write about this as metaphor. He is worried that that’s a place where we are heading. And he worries that he is already somewhat less human. On the other hand, I try to think back to what the lives of my ancestors 20,000 years ago were like--let alone the lives of my ancestors 80,000 years ago before the leap that was modern speech as we know it.
I think they probably spent a huge amount of time being hungry. I think they turned all of their processing power to trying to figure out the exact emotional state of everyone else of the 50 people they knew at any point in time. I think they were also trying to watch out for dangers, and for possible food sources. That is the natural life. And in some ways, we are still pretty close to it. Look at daytime TV. Look at the magazines in the checkout line at the supermarket. It’s pretty clear that at some fairly deep level we are still wired to focus on three and only three things:
- Possible dangers of violent death. especially as it affects children.
- Sources of food or perhaps of other valuable and especially rare. resources.
- Who is sleeping with whom--so you don’t get a black eye by making a bad mistake...
Brad DeLong: Well, why doesn’t it already know that when I am dictating I am like 10 billion times more likely to say the name of the Swedish economist "Knut Wicksell", than to say the phrase "nut whistle"? You’d think this would be easy...
Lance Knobel: Okay. We have time--I’m sorry--for just one more quick question. Let’s go in the middle there.
Male Speaker: Brad, you flirted with this question a couple of times during the panel. So if robots are going to take over a bunch of [0:31:04] [inaudible] work, there is a question how do we figure out how to pay the people who no longer have work. Would you be willing to hypothesis, do you have an answer that question?
Brad DeLong: Well, as I say so far, we have been very lucky that so far we have been granting people their social worth and also granting people their control over material resources depending on what they have been able to concretely add to our collective store of value--what they are able to make. And of course, we also have been assigning people their social worth and their control over resources by what they are able to convince us they out to be able to grab from the common store. Because by now an overwhelming proportion of all of our wealth is our common joint product rather than our individual product.
After all, put any of us naked out in the Sierra Nevada--even during the rainy season--and our chances of even surviving to make it out of the Sierra Nevada are pretty small. Practically all of what we do is a joint product. So far our system of resource and statu allocation worked extremely roughly. But it has been semi-tolerable. At least around here most people get enough to eat. You scratch your head on how it is that Mr. Fitzwilliam Darcy is able to consume £10,000/year of stuff, given that he appears to have no social utility at all--except to be occasionally witty but mostly cruel. And yet somehow he has convinced people that this amount of the common store out to be his.
We are moving forward into a world in which we are going to have to figure out a better way to share out of the common store. We won’t be able to rely on people’s ability to do something actually concretely physically useful in order to get control over the material resources and also to get the validation of self-worth we seem to need to make the system tolerable.
My friend and coauthor Larry Summers touched on this a year and a bit ago when he was here giving the Wildavski lecture. He was talking about the extraordinary decline in American labor force participation even among prime-aged males--that a surprisingly large chunk of our male population is now in the position where there is nothing that people can think of for them to do that is useful enough to cover the costs of making sure that they actually do it correctly, and don’t break the stuff and subtract value when they are supposed to be adding to it.
That is a problem that human societies have never faced before. Larry thinks it may be that we are starting to see the thin edge of the technological wedge in male labor force participation trends over the last 30 years--which I point out is not matched at all by female labor force participation trends. Women were even very briefly at a premium on the job market in the last recession--even given the extraordinary extra share of childcare and household production labor that they do.
It’s plain that the next time there is a recession, there will be more females than males working outside the home again. It's a worry. It’s a scary worry for our grandsons and great-grandsons, and perhaps for our great-great-great-granddaughers as well...