Supply Creates Its Own Demon: Marc Andreessen and "Textbook Luddism"
by Sandwichman at Econospeak
Supply Creates Its Own Demon: Marc Andreessen and “Textbook Luddism”
Marc Andreessen has a column in the Financial Times with the headline, “Robots will not eat the jobs but will unleash our creativity.” Here are the first two paragraphs:
A growing number of people seem to fear that robots will eat all the jobs. Their worry boils down to this: computers can increasingly replace human labour thus displacing jobs and creating unemployment. Your job, and every job, will go to a machine.
It is textbook Luddism, relying on a “lump of labour” fallacy – the idea that there is a fixed amount of work to be done in the world by humans. The counterargument comes from economists such as Milton Friedman, who believe that human wants and needs are infinite, which means there is always more to do.
Mr. Andreessen knows as much about Luddites and the lump–of–labour fallacy as I do about programming browsers. Ordinarily, it might suffice to cite the ONLY published scholarly articles on the history of the phony fallacy: “Why economists dislike a lump of labor” and “The’lump of labor’case against work-sharing: Populist fallacy or marginalist throwback?” both by a fellow named Tom Walker. But Andreessen’s timing has caught me in the midst of a research/writing project that attempts to make sense of what Joan Robinson identified as “mumpsimus“: the persistence of discredited arguments in the face of overwhelming evidence.
I’m about 20 pages along in my new piece, Supply Creates Its Own Demon. The demon in the title refers to Maxwell’s demon and, by association, the chess-playing automaton (an elaborate hoax) built in the late 18th century by Baron Ludwig von Kempelen. I’ve just gotten to the section where I discuss Andrew Ure’s 1835 The Philosophy of Manufactures. Ure’s book contains a discussion of automatons, which includes the chess-player but doesn’t mention its imposture.
The third section of Philosophy of Manufactures, “Moral Economy of the Factory System,” relies heavily on Edward Carleton Tufnell’s supplementary report for the Royal Commission on the Employment of Children in Factories, which is one of the most sustained anti-union diatribes in English literature. Tufnell went on to write Character, Object and Effects of Trades’ Unions. I have credited Tufnell’s diatribe with “putting legs” on the bogus fallacy claim — I could amend that to say Ure’s appropriation of Tufnell’s claim put wheels on it.
An article by Steve Edwards, “Factory and Fantasy in Andrew Ure,” makes a convincing case for the influence of Ure’s Utopian analysis of the factory on Marx’s analysis of “real subsumption of labor” in the originally unpublished “chapter six” of Capital, “The Results of the Immediate Process of Production.” Marx’s analysis is, in a sense, an “immanent critique” in that it uses insights from Ure’s text to highlight incongruities and contradictions in his argument.
Briefly, Ure argued both that workers delayed technological progress that would havebenefited them through collective action and that collective action by workersaccelerates technological advance, to the detriment of the workers. Or more simply: strikes delay and accelerate technological progress that helps and hurts workers. This is a “nice knock-down argument” to be sure but it couldn’t be more arbitrary.
Coming back to Marc Andreessen’s column, his argument is a pale shadow of Ure’s. For all its overt hostility to workers and unions, I prefer the original Philosophy of Manufactures because in its fantastic exposition it laid bare the essential incoherence of its premises.
Both can be true, in a specific sense. Is the argument intended to be global/absolutist?
“Both can be true, in a specific sense.”
Absolutely. (I presume you’re talking about Ure’s delay/accelerate benefit/detriment complex.) Both — even all four combinations — can indeed occur. But for that reason, they can’t all be “inexorable.” Ure was invoking “laws” that could brook no opposition — except, of course, his laws were in opposition to themselves.
Yes, that was specifically my question.
Not sure why any critical reader would take this opinion piece seriously, the first quoted paragraph immediately launches off as a straw man argument in multiple respects.
Wasn’t the chess fraud the Mechanical Turk? I believe Maxwell’s Demon was a thought experiment involving a hypothetical demon that would sit at the opening of a container, only allowing warmer, or colder, particles to enter the bottle.
However, to address the Lump-O’-Labour Fallacy, something left out fairly often is that while there may be plenty of labour to go around, cotton to pick and babies to tend, the amount of PAID labour is another matter.
This straw man has a deceptively benign appearance. When you look closely (historically) he is blowing a class-war dog whistle. Class War, not “class war.” Workers were being shot and hanged by the state at the behest of the factory owners at the same time they were being ridiculed for having “foolish ideas” that the capitalists and their apologists in the journals attributed to them.
The lump-of-labor straw man began as a condescending, paternalistic jibe at the intemperance of the poor and then mutated, with E. C. Tufnell, Nassau Senior, Andrew Ure and the cotton masters of Manchester into a full-blown calumny against trade unions specifically and collective action by workers in general. After the repeal of the Combinations Acts in 1824 and 1825, it was the advice of Senior to the Home Secretary to prosecute trade unions on trumped up technicalities, so that they would remain de facto “outlaws” even though unions were de jure legal.
The straw man doesn’t simply invoke a cartoon version of history, it invokes the suppression and enforced forgetting of a real history that was and remains vital to understanding the issues of technology and employment.
“Wasn’t the chess fraud the Mechanical Turk?”
Yes, Noni. It was a contraption built by Baron von Kempelen, who also made some genuine automatons. But the chess-player was a puppet operated by a person hidden inside the box where the mechanism was supposed to be.
In the piece I’m writing, I draw a “six degrees of separation” progression from von Kempelen’s ersatz automaton to Maxwell’s demon that proceeds from the chess-player through Edmund Cartwright (power loom), Peter Ewart, James Joule and Rudolf Clausius to James Clerk Maxwell and his demon. Watch for it!
I guess I would be one of those destroyers of Labor just by the work I do in throughput analysis, improving the flow of materials and labor input during production on the shop floor. But even with automation, there still is a limit as to what can be made in any particular increment of time. I ended up explaining to one Manager they had exceeded the capacity of a four cavity tool on some days and not on other days and were delinquent in delivery they had promised. No amount of Labor would have increased throughput. A partial solution would be to run full out on all days and hope the mold does not break. So much for checking capacity before making promises or accepting orders.
I think the larger question here is where do the productivity gains from throughput improvements go. You have stated it could come at less hours work and still being paid for a full day or week. As you well know, we are not seeing such and a limited amount of Labor is being used the same as the tool I described above. And much of Labor agrees with this scenario and damn the slackers.
The issue with the tool the same as Labor is experienced over a period of time. The tool wears and becomes less efficient as you use it 100% with no time for maintenance or adjustment. Feathering occurs or a pin breaks and greater damage occurs and the tool is taken down whether or not it can be or product is released to the field and must be recalled. The scenario exists the same as it did with Cobalt and Ion ignitions.
I read the FT article http://www.ft.com/intl/cms/s/0/fc1001e0-f888-11e3-815f-00144feabdc0.html?siteedition=intl#axzz35bLYOfjE . The message is disjointed. If costs go down due to productivity gains made by improved throughput sans Labor, where is the income to Labor from the accelerated production? 1% of the taxpaying households making >$500,000 is approximately 1 million people. ~7 million households make > than $180,000. The larger question is, if this is enough to sustain a economy where no one else can afford to buy the cheaper good due to lost income? Andreessen does not address this.
Unburdened Direct Labor is < 10% of the Cost of Manufacturing. Beating up labor is a thing of the past as Drucker would put it. So why bother with robots if no one can afford the cost of cheap goods once they are in the Not In Labor Force?
– “offer better access to education and skills development, which itself will increasingly be delivered through technology, to help them train for or even create new jobs.” Have them sign for those student loans so they can become indentured to the federal gov or the banks for a hundred thousand or more for a masters.
– let markets work, this means employment contracts that are easier for both employees and employers to terminate – so that capital and labour can be reallocated rapidly to create new fields and jobs. Markets already work in this fashion in that we have a larger than normal Labor pool. Unless someone is up to moving to China or Asia, people are not going to move.
– A secured safety net? How about paying me for the productivity gains from the improved throughput? Isn’t this what is supposed to be happening with productivity gains? A portion to labor and a portion to capital.
Anyway my $.02. I love talking to these guys as I do this already on the shop floor.
In Catholic high school religion class they taught us that we were made to possess the infinite — God — so that no matter how much we come to possess we will always want more. Just want you to know the Catholic Church is behind you, Sandwichman.
The thing is that the Luddites were absolutely correct. The workers got nothing from the Industrial Revolution until the Chartists started smashing things. Workers had nothing to gain from the factory system, so as rational economic actors, they should be expected to oppose it. Collective action, and violent collective action at that, was the only thing that did work. Arguing otherwise ignores history as this pattern is repeated again and again.
Besides, Andreeson knows nothing about economics or business. He’s a government welfare baby. Didn’t he make a pile of money on some NCSA government funded piece of software that operated on the government funded internet. If I want serious business or economic opinions, I’ll ask someone with some experience in the area.
Yes, the “textbook Luddite” fable is a libel. It’s the same old story of taking people’s land, disrupting their lives, enslaving them and then calling them “savages” when they resist.
Judging by the title of your “new piece”, it appears that you have homed in on Say’s law as the foundation of a lot of the nonsense spouted by economists over at least the last 35+ years. (Supply Creates Its Own Demand)
Apparently Say never foresaw that economics would become more like religion than science. How very short sighted of him. I guess that in his world, theories were expected to be modified to comply with the perceived reality. Instead his theory became law. (Ignoring the assumptions that it was built on.)
But having perfected economic theory by 1980, all that was left was to apply it. (And explain away the observable contradictions.) Thus Say’s law led us to reduce taxes on producers, and to allow producers to out source production to foreign countries with minimal tariffs on import. Because every good economist understood that if you built it then it would be purchased. Because Say’s law said it and that was that.
And once a nation allows or encourages wholesale out sourcing of production to foreign countries then why would it resist the wholesale domestic automation of production as that out sourcing of production was reduced?
The answer should lie in the result of all that initial out sourcing of production. That result being huge amounts of consumer debt accumulated to maintain a standard of living which the consumer had come to expect. And in the later stages, consumer debt grew as students faced rising costs because revenue starved government, reduced spending on education. And in our current stage, consumers face the need to replace an automobile that is aging past any former expectation of use and that will raise debt further or reduce spending on other goods. That total consumer debt reached $12.68Trillion and seems to have stabilized a little below that peak. Now standards of living will have to fall for the majority of Americans. Especially for those making less than the median wage.
So does said nation draw a lesson from this or does it continue to the second wave of destruction. (i.e. automation)
That depends on how much of the economic mumbo jumbo will continue to be swallowed by our elected representatives. Always remembering that their campaign contributions depend on their swallowing it.
It has only been 6 years since said nation got its comeuppance with a ginormous recession. And judging by the unemployment and labor participation rates, it is continuing.
It may take us another 15 or 20 years before the swallowing ends.
Japan has again cobbled up some inflation and thus they assume their economy will take off. Apparently the swallowing is not done there and they have been at this much longer than we have. I would have thought that sooner or later someone would suggest attacking the cause of their problem instead of an effect of their problem.
Consumers can not spend what they do not have. Producers will not produce what they can not sell. Where does this end.
“If you worker bees won’t embrace the way I insist on devaluing and unskilling your meager livelihoods I’m going to to keep making fun of you in the media until you cheer up.”
JimH “… it appears that you have homed in on Say’s law as the foundation of a lot of the nonsense spouted by economists over at least the last 35+ years.”
Yep. Say gets a bum rap in all this. It wasn’t his “law” and he wasn’t as dogmatic about it as the vulgar epigones. The story, as I see it, is more that capital and its apologists have what Charles Tilly called a “repertoire” and “supply creates its own demand” is a slogan that affirms the legitimacy of that repertoire. To return to my central metaphor, Say’s Law invokes the “automaton” but it is actual people inside the ostensive mechanism that are playing the game and pulling the levers.