Vernor Vinge is consistently one of the most interesting and conceptually dense futurists I’ve had an opportunity to listen to. While watching this excellent talk of his at Singularity University, my ears perked up at the mention of technological unemployment, the primary focus of this blog.
About halfway into the talk he broached the general issue of technological disruption:
“In the present era we all seem to be involved in the Red Queen’s race. Myself as a writer, I’m up against eBooks, and I’m up against all the piracy. I’m racing as fast I can, and if I hadn’t actually had some success in the past, it would be of course much, much worse for me to be in this steam turbine that’s called modern progress. When I talk about it happening to some other job category, I don’t feel quite so tragedy-struck about their plight. But we’re kind of all up against a situation of terribly disruptive new technologies.”
Soon afterward he addressed the topic of technological unemployment more directly:
“What comes after technological unemployment? There are certain things that humans are still very good at. And some of these crowd-sourcing [successes] tell us that those certain things that humans are still very good at are maybe much larger than we think in comparison to the machines, at least if the work is done in coordination with the machines. But what we’re really good at are isolated things. So you can imagine a civilization in which there are these bright little sparks of human level intuition and creativity and insight that are separated by vast stretches of algorithmically accessible problems. And there’s a lot of occupations and businesses, where the successful insight on the part of management is figuring out how to do all the stuff you can do without those expensive people, and then what remains are those bright little spots where you need to have the people.
So technological unemployment, I think, is a very real thing and, in the early twentieth century, the white collar and academic sorts could kind of say grandly, and with great objectivity, ‘Don’t worry about the buggy whips no longer being something that can be sold, there are other jobs that technology will create for you.’ I think their objectivity is cracking a little now that the tide of automation has risen to the point that more and more of their academic and white collar jobs are also under that sort of pressure…
Having a survival strategy for that sort of world is one that might get us automatically into a situation where, in the late teens, the average form of at least white collar work throughout the world is actually to be participating in social networks or group minds that allow enterprise to focus the humans on the things that the humans can really do and can do much better than any contemporary machine. I mentioned at the beginning of the talk that in past years I kind of turned up my nose at the human thing in the long run, because as I said ‘biology doesn’t have legs.’ However I think it’s true that at the present time, there’s no machine on earth that is of the intellectual caliber of a human being. Personally I think that we will get that and things that are much better. But the present situation is that we don’t have anything like that. On the other hand, we already have an installed base of seven billion human equivalent intelligences out there, and they’re self sustaining for the most part. That is really a remarkable resource, and this talk could be regarded as talking about how to make that resource all that it can be, for the benefit of humanity.“
Essentially Vinge is describing a theoretical crowdsourcing platform (or platforms) that could systematically harness human minds and direct them towards the tasks humans still do best. This is actually very consistent with the point of view expressed in Brynjolfsson and McAfee’s book Race Against the Machine, in which unemployed human beings are described as a large “slack resource,” one that hopefully some entrepreneur or platform designer will figure out how to put to efficient use. I do think this is a very credible possibility and I particularly like Vinge’s image of a few bright shining stars of human ability twinkling against the blackness of algorithmic space. True, over time, more and more of those stars might blink out, and new stars might become much harder to find, but it’s entirely possible that there are still enough stars out there to see us through the coming transitional period, and computer networks might have the potential to help us identify and exploit those stars.
Of course today we do not have anything like such a platform, though we might have potential precursors. Vinge’s examples from the talk, such as Wikipedia and Foldit, have been tremendous successes, but they do not compensate the people involved. Alternately many of Brynjolfsson and McAfee’s examples, such as eBay and Threadless, are merely retail marketplaces and don’t seem particularly future proof to me. Taskrabbit and Mechanical Turk seem a bit more promising, though neither works on the scale or level of efficiency that Vinge seems to be implying might be possible or necessary. It will be fascinating to see what entrepreneurs come up with over the next ten years.
During the Q&A section, Vinge was asked (by Federico Pistono, interestingly enough) to elaborate on technological unemployment and how we might transition to a post-scarcity economy. Vinge began by addressing issues with the term “post scarcity”:
“Let me start out with a small disagreement in terms of terminology. A “post scarcity” economy is certainly not a post singularity idea. I think that the reach of the mind will always exceed its grasp. Whatever is thinking will always be able to think of projects that are beyond what it can presently do. And that automatically gets you into a sort of scarcity situation. So sometimes when I read utopias, where everyone has everything they want and that’s the start-off assumption, that’s something I disagree with.”
While this is merely a semantic point, it is one I agree wholeheartedly with and tried to express in my own way in a recent blog post. As long as some mind somewhere wants something, then scarcity will probably exist. However, in Federico’s defense I believe that he was using post scarcity more in the sense of “post material scarcity”— as in a society where everyone’s basic needs are met and people have the ability to opt out of wage labor and the market economy if they so desire.
Vinge then continues:
“On the other hand, there’s still the very valid point about what happens when large numbers of people cannot economically compete with what they do and that there is no alternative project on tap. I don’t call that post scarcity, but I do see it as a problem. Up until about six weeks ago I was kind of in a tailspin about this issue. But then I started preparing this talk, and it occurred to me that we still have a lot of mileage left, if you take advantage of the fact that we no longer have to have physical location involved in our work—and I mean that in a much stricter way than just telecommuting. At that point, the fact that there’s all these billions of human level intelligences out there is very powerful, and if you’re willing to work in that way, I think there is plenty of career left for any such person. That takes us a long way, and I think eventually I’m very much in favor of the [intelligence amplification] stuff. I think we will get into a situation where those people who prefer to work, in addition to having the social networking and group mind stuff, will have real group minds (that is where you have connections at the mind level) and intelligence amplification.
“Let me say a little bit more about real group minds. It’s scary in its own way because you come face to face with issues of identity and awareness. I’m gonna get mystical here for a minute. Imagine that the human soul is like processes not just on a single machine operating system, but on a network. In that case, depending on the bandwidth and depending on the size of the problem, the size of the mind that is working on it changes. In that sort of world, you’d have super intelligences for a few minutes and then they would decrease perhaps into isolated intelligences if there was some reason for working in that way. So there are some very strange things that are gonna happen in the latter era as we get into the singularity and I don’t claim to know what they are. What I just went through is some dreaming about how strange it could be. I think there is full employment in a situation like that, and our goal as present day group minds is to make sure full employment is entertaining to us.
“I think that especially if things work out in a mellow way that there will be significant number of people who choose to be consumers but not participants. And that the people who make things, and the automation that makes things will be plenty powerful enough to satisfy those consumer desires.”
Later, in response to another question, Vinge continues along these fairly optimistic lines:
“About five or six years ago I gave a talk at a corporate research center and I brought up the issue of the internet as a testbed for group enterprise involving large numbers of people and someone raised their hand and said, “The average person, do you really think they have either the motivation or the talent to be of use?” And I think I can give a somewhat better answer now than I did in that case. The short answer is: yes. And then caveats: I think it was Voltaire who said the average person just wants to cultivate their garden,’they’re interested in their friends and their family, and they don’t want someone to come and blow it all away. But that still is a human level intelligence. If it turns out that a group mind or social network sort of employment is still something that can pay them money, they may very well be willing to donate their human level intelligence to participating in social networks to solve problems and get money for it. Then also if you look at the demography—this is intuition now, off the top of my head—that I see in contributors to Wikipedia, I think that there is an enormous number of people—perhaps still a small minority compared to the total human population—that have intellectual interests of a certain sort and certain sorts of talents, and they really think it’s cool to work with that and get that out to where it can be used…
“This is very likely going to be the sustainable form of employment for the next fifteen or twenty years, and I think that gives them plenty of encouragement to do something that otherwise is very pleasurable to begin with. So I do think the human race as a whole has these segments that can happily participate in doing very clever things.”
Of course there are a lot of unresolved questions regarding Vinge’s vision. It’s not clear that the marketplace will actually compensate people who participate in these crowdsourcing/social networking/group mind schemes. As Vinge points out, people participate in these tasks for a variety of reasons, often because they enjoy it (Wikipedia), or because the designer of the platform has managed to make the task into a game (Foldit). Making sure people get compensated for the crowdsourcing work they do might require some sort of government mechanism. It is unclear at this point.
However, as far as the general notion that we might “still have a lot of mileage left” in terms of useful tasks for humans to perform, I think that position does have a fair amount of plausibility.
In their book 
A lot of more traditional pundits admit there is a potential problem with technological unemployment (or at least technological inequality) but seem uninterested in any of the more drastic solutions mentioned above. Instead they push for a series of common sense policy fixes, such as fixing education to better prepare people for STEM fields or reforming the patent system to mitigate drags on innovation. These policy tweaks are designed to make the economy function incrementally better in this new technological era.
One line of reasoning argues that technological unemployment can only ever be a temporary problem, since the day is approaching when we will literally print out all of our food and other necessities. Thus the problem only arises in the awkward transitional period: after we’ve automated large numbers of jobs, but before technology has lowered the cost of living to near zero. Therefore we should try to accelerate technological progress by whatever means necessary so that we can make the painful transition as short as possible—much like tearing off a bandaid.
This position argues that as long as government stays out of the way the transition will be relatively smooth and painless. Yes, there will be less jobs available, and certainly people’s incomes will suffer, but technology will simultaneously bring down the cost of living at a fast enough rate that people will survive just fine without the need for government invention or economic restructuring.
It’s possible a technological unemployment crisis would simply be a pitstop on the way to a much more dangerous crisis involving artificial general intelligence. Once AGI arrives we will have much bigger issues to contend with, such as will the human race survive being displaced as the most intelligent beings on planet Earth? So while technological unemployment might be worth worrying about at the margins, really the bulk of our energy ought to be devoted to this more substantial threat on the horizon. This position is advocated by the 