June 18, 2015
Robot Impact On Productivity
Lately robots have been boosting productivity about 0.35% annually. By comparison, IT has been boosting productivity by about 0.60% annually with about 5 times as much capital expenditure. So higher ROI from robots? Keep in mind these numbers could contain substantial inaccuracies.
Does it make sense that IT is doing more than robots to boost productivity? My intuition is yes because so far robots have been used mainly in manufacturing and manufacturing is a much small portion of the economy than it used to be. Manufacturing only makes up 8.8% of total employment in the United States, about 12 million workers in 2013. So robots need to spread out into a lot more parts of the economy in order to have a large impact on productivity.
Check out this table of employment by sector in the US economy. 5 different sectors each employ more people than manufacturing does. Consider the prospects for automation in each of these sectors. For example, the retail trade has about 15 million worker, wholesale has about 5 million, and transportation and warehousing has about 5 million. Well, picture self-driving trucks, self-driving local delivery vehicles, totally automated warehouses, and easy online ordering. The need for store employees, vehicle operators, and warehouse and wholesale employees will drop dramatically from robots and IT systems. Goods will move from factories to end purchasers with little human involvement.
I expect IT systems will continue to automate a lot of financial work including investment management. Index funds beat active managers. Robo-advisors for money management will probably make it easier to manage your money. Betterment, WealthFront, Charles Schwab, and Vanguard are among those offering automated investment advising.
In what occupations will robots start to make major in-roads in the next 5 years? Do you know about robots that are under development for a purpose in your industry? Where can you see them fitting in outside factotries?
Randall Parker, 2015 June 18 07:37 PM
Robo tractors in Australia, Europe and Japan and then later in the United States. (Cheap farm labour in the US will retard the uptake of agricultural robots there.) All the technology has been worked out for driverless cars without the need for the tractors to be able to drive in traffic. Farms are typically large enough to support at least one full time robo tractor in Australia. Smaller average farm size in the Europe and Japan may result in rental arrangements with the capital being worked 24 hours a day during busy periods.
Computer programming is still not advanced enough, constraining the operation of robots. Major breakthroughs in software will take a few more decades.
In addition, mechanical engineering, which is fundamental to the design of robots, is also very dependent on software.
Separately, besides software, the main food that robots need is energy. In the future, robots will be very hungry for energy, and the electric grid capacity will have to increase by an order of magnitude, which means that nuclear energy will have to be developed further, solar and wind power will not be enough. Thorium molten salt reactors will be key.
But once software and cheap energy issues are resolved, employment will no longer a problem because if the robots can increase production by a factor of 100, then even if 90 % of the population becomes unemployed, this will not upset the economy!
Wolfdog, very roughly a human might output 100 watts of work. So for a humanoid robot to do what a human worker does in a day will only take about 1 kilowatt-hour. If everyone in the world has the robotic equivalent of 2 human servants working for them, that will increase electricity consumption by about 2 kilowatt-hours a day per person, which is much less than an order of magnitude. Here about 2.5 square meters of solar panels could supply that much electricity, or two of the local 3 megawatt wind turbines could power robots for almost everyone in Adelaide. Now you may be thinking of a future where everyone has thousands of robots at there beck and call, but it does not take much energy to replace all human labour many times over.
If you don't see economic effects on society from 90% unemployment, then you have replaced humans with robots, too.
Even assuming ideological rapprochement allows the less democratic-socialist capitalst states to move to guaranteed income and other changes required for such a world, capital accumulations that don't require human involvement (other than the owners, of course) for the pursuit of various aims, adaption to a a life without mandatory work attendance for the majority, and a growing inequality gap are going to be massive society wide changes, probably on the order of the industrial revolution.
If ideology does become messy (in my view, p=>.99), well, you don't dislocate 9/10s of the participants of a rich economy without spilling your drink.
Utopians, it seems to me, need to explain how humans get past the many popular zero and negative sum games they like to play.
Why does everyone assume the robots are coming, the Singularity is coming? Consider this. Over the last 30 years or so, the real income of the working class has declined sharply, and the real income of the middle class has declined modestly. The rich have gotten richer. Also, work force participation is declining, partly due to retirements, but also due to lack of jobs. And many workers have been forced into lower paying or part-time jobs. All this suggests that real GDP is increasing more slowly than population, if not actually falling, and that there is less stuff to go around.
The productive, white population has stopped having babies, birth rates are below replacement level everywhere, and in some countries the white population is in actual decline. The same thing is happening in China, Korea, Japan, Taiwan, Singapore, et al. There is also a great deal of abandonned things, amusement parks, factories, almost whole cities (Detroit, eg). In Europe, there are many abandonned villages. Wolves roam some abandonned towns in East Germany and hunt deer. Europe and America are being overrun by savages.
So, why do you think the arrow is pointing up? Aren't we in the midst of a slow motion civilizational collapse? Did the Europeans of the 5th Century see a sudden collapse or a slow decline? All this talk about the coming robot revolution is delusional.
I assume the singularity is coming, and I assume the government has decided that income inequality is useful, and to be encouraged. Because it concentrates wealth and power, making them easier to extract/control.
I mean, you're a dairy farmer, which do you want, a cow, or a million mice? We're the mice, billionares are the cows. The government is the dairy farmer.
Automation is increasing the productivity of the economy, but by a variety of means, the increased production is being centralized.
Gonna be really nasty when they close the loop on automation, have lights out factories, and don't need the mice at all.
Ronald Brak: "Now you may be thinking of a future where everyone has thousands of robots at there beck and call, but it does not take much energy to replace all human labour many times over."
Humanoid robots who vacuum the floor don't count as robots: real robots are totally automated factories and mining machines that consume a lot of energy. In fact, in the future, as production becomes more ambitious, the amount of energy consumed by undersea mining operations might require 10 times the energy that the world is consuming at this moment. Giant automatic factories that are modularized and self-modifying will count as robots.
J R "Bob" Dobbs said at June 19, 2015 6:03 PM:
"Wolfdog,If you don't see economic effects on society from 90% unemployment, then you have replaced humans with robots, too."
/* --------------------------------------------------------------------------- */
The 90 % unemployment will happen regardless of which political system we adopt because automation will keep making progress. But my point was that even if unemployment becomes 99 %, the improvements in automation will increase productivity to the point that even only a small fraction of the robots will have enough spare capacity to take care of the 99 % unemployed without offending the rich. The only obstacle to the age of abundance will be if new sources of nuclear energy are not developed to feed the robotic factories and mines.
And contrary to what you said, the unemployed 90-99 % will not become robots: they will have more time to study music, philosophy, science and engineering at their own pace, instead of doing unproductive redundant jobs that are actually unhealthy and humiliating like pumping gas or flipping hamburgers. Food, shelter and basic health care can be affordable and plentiful for all average citizens if robot capacity increases, provided that new energy sources are developed to feed the robots. Then the so-called unemployed will have the freedom to do more human activities like teaching philosophy to children or to take talk to elderly patients in the hospital.
I think that the answer to the question of whether changes in technology will result in mass unemployment will in part depend upon the ability of governments to tax the increased wealth/productivity.
If governments can find ways to effectively tax the gains accruing to the owners of the robots/technology then they will be able to afford to employ a lot of people (e.g. in education, health, military, arts/entertainment, etc). A Government's ability to employ people is limited only by the amount of money it can collect without destroying the economy. A government doesn't need to be efficient in its use of people like the private sector.
Why I think the robots are coming:
- robot sales are rapidly rising.
- computers and robots are getting more powerful every year.
- robots do an increasing list of tasks better than humans. So not just cheaper but better products.
- robots do not go on strike, get sick, quit, or laze off.
- robots can be moved to other factories in other countries.
- robots can be upgraded quickly.
Abandoned places: There are also many growing places. There are places where the brains are congregating and producing lots of innovations and product designs and new services.
Robots that vacuum the floor count because they replace human labor. Most manual labor is going to get replaced by robots.
Bob Sykes, do you think only white people are productive? That seems an odd thing to write. How are you measuring productivity? In terms of the rate of output per unit of input? If so I don't think there's any reason to think that what's generally thought of as "white" people are any more productive than non-white people. Are you thinking in terms of economic growth? Because Africa as a whole, the population of which is generally considered to be non-white, now averages more than twice the economic growth of the United States. Is it average hours of work per worker? Because there the US averages higher than the "whiter" members of the OECD, but Mexico, a country I would guess is mostly considered to be non-white, is way out in front with an average of 2,237 hours of work a year per worker compared to the US's 1,788. You are right that real wages for poorest quintile of American workers have stagnated or gone backwards over the past couple of decades while they have steadily increase over that time for the poorest quintile of workers here in Australia. This is due to capital taking an increasingly large share of the economic pie, and I think this situation has partly come about due to racism in the US which is used to convince people to vote against their own economic interest.