Technology has a way of dazzling us into the deterministic fallacy: assuming path dependence for the ways a technology develops and its impact on society. But we have agency.
The so-called “productivity bandwagon” that we assume follows a new technology (where Schumpeter’s creative destruction will generate more jobs than are destroyed) is not inevitable. Widespread gains require that a technology creates more demand for workers (by creating new tasks and industries), and that demand induces higher wages. Neither are certainties, and take societal negotiation between labor and capital.
That’s from a 2023 book co-authored by economists Simon Johnson and Daron Acemoglu called “Power and Progress: Our Thousand-Year Struggle Over Technology and Prosperity.”
The book is big and thorough, with a sprawling historical comparisons over a millennia. Yet, I was disappointed by how few concrete examples the authors gave for what precisely they want to be done differently — especially in a book that is more than 500 pages. For example, on page 353, they write “digital technologies, which are almost by their nature highly general purpose, could’ve been used to further machine usefulness – for example, by creating new worker tasks or new platforms that multiplied human capabilities.” But that “for example” is not actually an example, but rather a reassertion of the general outcomes they seek (“new worker tasks or new platforms.”) Instead, I wanted an example of what exactly could have been done differently to ensure new worker tasks or platforms.
In that way, I found so big a book disappointing, and felt it could have been half as long. I appreciated their overall point, though, of idealizing “machine usefulness” in four ways: machines should improve worker productivity; create new tasks; distribute accurate information (like the web) and give better access and markets. Just don’t look to this book for the path to get there.
As they write: “How technology is used is always intertwined with the vision and interests of those who hold power.” Below I share my notes from the book for future reference.
My notes:
- Bentham’s 1791 panopticon informed Foucault’s philosophy and Orwell’s 1984
- Thousand years of progress: “shared prosperity emerged because, and only when, the direction of technological advances and societies approach to dividing the gains were pushed away from arrangements that primarily served and narrow elite. We are beneficiaries of progress, mainly because our predecessors made that progress work for more people.”
- 18th century John Thelwall of monopolies: “All diseases not absolutely mortal, carry, in their own enormity, the seeds of cure” Workers congregated in cities and organized. This was an outcome not the purpose of industrialization
- Keynes: “technological unemployment” (“ This means unemployment due to our discovery of means of economizing the use of labor out-running the pace at which we can find new uses for .”) and Mary Shelley’s Frankenstein were early warnings
- In 1821, David Ricardo updated his opinion noting there were risks of technology displacing labor
- But Francis Bacon’s 1620 Novum Organum out technology enthusiasm
- Authors note declining male labor force participation, the focus of the book Men Without Work)
- “There is nothing automatic about new technologies bringing widespread prosperity. Whether they do or not is an economic, social and political choice.” (13)
- “Productivity bandwagon” is conventional wisdom, that more productive sectors bid up wages for those outside of it. That is not inevitable
- “To clarify the distinction between output per worker and marginal productivity, consider this often-repeated prediction: “The factory of the future will have only two employees, a man and a dog. The man will be there to feed the dog. The dog will be there to keep the man from touching the equipment.” This imagined factory could churn out a lot of output, so average productivity — its output divided by the one (human) employee — is very high. Yet worker marginal productivity is minuscule; the sole employee is there to feed the dog, and the implication is that both the dog and the employee could be let go without much reduction in output. Better machinery might further increase output per worker, but it is reasonable to expect that this factory would not rush to hire more workers and their dogs, or increase the pay of its lonely employee.”
- Average productivity vs marginal productivity (of adding one more worker)
- Different kinds of automation: self-check kiosks displace scanning from employees to customers, it doesn’t unleash new productivity or make food cheaper (what they call “so so automation”)
- Productivity bandwagon requires (1) more demand for workers and (2) that demand Induces higher wages (says neither are inevitable)
- Mark Zuckerberg’s 2018 announcement from a few executives: prioritize “meaningful social interaction” between people (rather than news orgs). This wasn’t the inevitable step but one chosen and then shaped follow on efforts (authors compare it to the Chinese Communist Party rolling out its social credit system)
- General purpose technology in particular: once a path is started it starts feeling inevitable direction
- The “vision oligarchy” of tech leaders
- Different kinds of technology: creating new tasks and opportunities or a preoccupation with cutting labor costs
- French master De Lesseps was over confident about his Panama Canal plan after his Suez success — as David McCullough’s classic 1977 book reminds — and so his innovation was not challenged. Tens of thousands died.
- Napoleon took Egypt with French Republic plans for a Suez Canal but got distracted for a century
- French canal in Toulouse in 1620s and the Erie Canal in the 1820s were seen as triumphs of engineering, science and technology for industry
- Mohommad Ali in 1811, Ottomans controlled Egypt after French withdrawal in 1801, hosted elite Marmuluks for a dinner where he had them shot
- Lesepps believed that “Governments can encourage such enterprises; they cannot execute them. It’s the public then on whom we must call…” believed industry should own and operate canals
- Lesepps: “a ship must not now be delayed”
- “Visionaries derive their power partly from the blinders that they have on – including the suffering that they ignore.”
- Lesepps always demanded sea level canal (no locks) and Suez proved successful, right as faster steamships took hold of international trade
- Research showing kids will imitate an unnecessary step to open a plastic box (but chimpanzees won’t) because they feel the adult had prestige. Likewise other research shows social media posts with lots of engagement labeled misinformation still stay with people
- The persuasion power of elite bankers framed the policy debate during Great Recession as options between bailing out firms or letting them fail — but ignored so many other possible options, like limiting bonuses paid out or coupling with homeowner protections
- Famous Lord Acton letter: “Power tends to corrupt and absolute power corrupts absolutely. Great men are almost always bad men, even when they exercise influence and not authority: still more when you superadd the tendency or the certainty of corruption by authority. There is no worse heresy than that the office sanctifies the holder of it.”
- Dacher Keltner research reinforces the point that power corrupts (his “power paradox”: Even good people turn less good with power)
- Marquis de Condorcet jury theorem for democracy
- Francesco Pertrarch introduced the myth of the dark ages, but science / technology advanced, including animal manure fertilizer, better rotation of crops, horseshoes, heavy wheeled plow and wheelbarrow, mirrors, steel and coal advances, early chimneys and mills (water and wind) but these gains did not lift European poverty. Where did the surplus output go? Cathedrals
- As much as 20% of total French output between 1100-1250 went to religious buildings
- In 1086 England, the church owned 1/3 agricultural land tax free and king held 1/6 of all land by value
- Peasants revolt of 1381 ended when Richard II promised to end serfdom but then renegged
- Medieval society was a society of orders in which those who fought gave special privilege to those who prayed so they would help control those who worked (the land)
- “How technology is used is always intertwined with the vision and interests of those who hold power”
- Malthusian Trap: authors argue elites coerced this trap; it didn’t occur always (periods of Greek and Roman heights had hundreds of years of growth without the trap)
- Statutes of Laborers of 1351 attempted to curtail post Black Death wage premium but could not limit growing wages
- “A House of Commons petition of 1376 put the blame squarely on how labor shortages empowered servants and laborers, who “as soon as their masters accuse them of bad service, or wish to pay them for their labour according to the form of the statutes … take flight and suddenly leave their employment in district.” The problem was that “they are taken into service immediately in new places, at such dear wages that example and encouragement is afforded to all servants to depart into fresh places…” 113
- Calorie-dense grain cultivation was another technology that resulted in less equal society: elites emerged in Uruk and Memphis with much bigger societies (This theme was identified in Dawn of Everything book too)
- Foragers lived longer and ate more diversely than their new farming descendants
- 1773 Enclosures Act makes it easier for large landowners to claim what was once common land peasants could farm on (elites see as unproductive use of land so they take it back and benefit)
- Authors tell story of Arthur Young who started his career as an enclosures defender but by early 1800s he had seen data and examples of worsening English rural poverty to flip his opinion to defend common space
- TS Ashton: About 1769, wave of gadgets swept over England”
- Daniel Defoe’s “Projecting Age” around the end of the 17th century, when western Europe was full of ambitious projects
- Authors use George Stephenson as an example of Defoe’s Projecting Age: rising through the ranks in a way uncommon during Middle Ages
- Industrial Revolution, why Britain first? (Why not China or Italy or Germany or France?) Authors group (and critiques) common suggestions into five categories: geography; culture (including religion); natural resources; economic factors (like wages and finance ) and government policies
- Authors reference (p. 159) industrialist Matthias Baldwin (1795-1866), whose grew one of the world s largest locomotive companies in Philadelphia, as an example that railroad technology wasn’t purely Britain strength
- Culture: Galileo had Catholicism to confront and though Newton had Protestantism after 1500s it was more lenient and less powerful
- Coal was near iron ore but the first Industrial Revolution phase was water powered textile factories (coal and iron more relevant by 1830s; first phase relied on cotton which wasn’t actually from Britain)
- Wages were higher in Britain but not by much over France and the Netherlands, and the productivity gains were so great that the marginal difference in wages seem unlikely to be responsible
- Glorious Revolution of 1688 set up strong parliament but France also protected property rights at this time; Britain did not earn most of the slave trade profits, and Britain didn’t encourage work policy anymore than anyone else — it was actually France with Jean-Baptiste Colbert who did most modern economic policy, and other European powers did more hands off (lassaiz-faire) government policy than Britain so neither less nor more government seems to signal
- What was the difference for Britain? Culture. “Nowhere else in the world at that time do we see so many middle-class people trying to pierce through the existing social hierarchy. It was these meddling sort of men who were critical for the innovations in the introduction of new technologies throughout much of the 18th and 19th centuries in Britain.”
- This was Defoe’s projecting age: Abraham Darby, Richard Arkwright, Josiah Wedgwood, James Watt couldn’t read Latin and weren’t reviewing scholarly works but trying to invent products
- The South Sea bubble of 1720 was itself an example of this pursuit of profit
- Rainhill Trials of 1829 in Liverpool set many global railroad standards
- Of 226 people who founded large enterprises during the industrial age, only two came from peerage and less than 10% had any connection to upper classes. But critically they also weren’t from the bottom, having fathers who had engaged in some industry and practical skills that they built on
- In 1577, clergyman William Harrison wrote in Description of England: “We in England divide our people commonlie into foure sorts” gentlmen/noboligy; citizens in English towns, yoeman farmers and then laborers
- This hierarchy barely changed for hundreds of years, when Gregory King in 1688 wrote Ranks, Degrees and Titles it was practically the same
- But English Civil War (1642-1641) and Glorious Revolution (1688) cracked open slowly a middle merchant class that could buy status with wealth and land, so they did so but did not improve the lot of their workers. (They wanted to join the rich, not change them)
- 1842 Report on the Sanitary Condition of the Labouring Population of Great Britain: English Victorians saw that industrial age and technology made some rich but worsened many other lives (real wages declined for most workers)
- “Technology’s bias against working people is always a choice, not an inevitable side effect of progress.” (179)
- Jan de Vries, the Industrial Revolution was an “industrious revolution” meaning we all worked more — in the 1750s, average English working year was 2,760 hours (53 hours a week), by 1800 it was 3,115 hours (60 hours a week): whether voluntary to pursue money, or coerced is a debate among historians and economists
- “The Luddites themselves seem to have understood not just what the machines of the age meant for them but also that this was a choice about how to use technology and for whose benefit.”
- In the words of a Glasgow weaver: “The theorists in political economy attach more importance to the aggregate accumulation of wealth and power than to the manner of its diffusion, or its effects on the interior of society. The manufacturer possessed of capital, and the inventor of a new machine, study only how to turn them to their own profit and advantage.”
- Why did first phase of Industrial Revolution worsen lives and impoverish and then by 1840s, see wages and living standard grow? In part it was the latter part featured inventions that didn’t automate away jobs (like spinners/weavers lost jobs) but gave new industries like railroads and telegraph
- US industrial technology boosted their relatively abundant unskilled labor, whereas many UK industrial inventions replaced their relatively scarce unskilled laborforce (199)
- Eli Whitney’s interoperable parts let more work on machines (American System of Manufacturing)
- Building labor power to 1871 trade unions (trade union act, gotten there with the lobbying of the Chartists for the Royal Commission on Trade Unions of 1867)
- Density of cities and factories helped people organize. John Thelwall: “a sort of Socratic spirit will necessarily grow up, wherever large bodies of men assemble” who will bring wisdom for power
- Edwin Chadwick’s 1842 report on poor conditions made many recommendations like running water into all homes
- Magus Alexander: “productivity creates purchasing power”
- 1938 Swedish meeting brokering the social democratic system, which included compromise between labor and employers — including that a company could boost efficiency over competitors and keep that profit. (American style welfare capitalism was more contingent on those companies donating)
- Authors speak kindly of FDR; but what’s the standard FDR criticism?
- Why were the decades after WW2 so good for all (Great Compression)? Authors say it’s because technology created jobs and workers shared the prosperity
- Authors argue Wagner Act and trade unions ensure direction of technology was productivity and not displacement; UAW and GE worker strikes immediately after WWW2
- Authors say M.I.T. Tech Square building ninth floor in 1959-1960 the beginning of the computer revolution
- 1970s northern California developed next stage with Lee Felsenstein and Ted Nelson (who referred to IBM and other controlling entities over compute as “cybercrud” in his Computer Lib manifesto
- Grace Hopper formed decentralized policies that became COBOL but….
- “digital technologies became the graveyard of shared prosperity.” — along with globalization and declining labor
- Authors say : 90% of children born in 1940 earned more than their parents did in inflation, adjusted terms, but only half of children born in 1984 did (true?)
- With digital tools, worker productivity went up but not worker marginal productivity
- “Technology does not have a pre-ordained direction, and nothing about it is inevitable. Technology has increased inequality, largely because of the choices that companies and other powerful actors have made.” (263)
- So how did business become so powerful relative to labor and why did technology turn anti labor? (264)
- The post-war market skepticism swung back to market prioritization by the 1970s after Buckley’s National Review (founded 1955) and Business Roundtable (1972, via merger) and Hayek informed Stigler and Milton Friedman (his books) rose in prominence with his doctrine (and the Jensen Amendment of tying compensation to share price
- Research shows CEOs with MBAs in U.S. and Denmark share less of company value creation with workers than CEOs from other backgrounds
- Coolidge: “The business of America is business;” but this became less popular after WW2, and then came back into vogue after oil price shocks of 1973
- Arrow replacement theory popularized by Christiansen as innovators dilemma
- Louie Brandeis: “We may have democracy, or we may have wealth concentrated in the hands of a few, but we can’t have both.”
- Robert Bork informed by Stigler set the standards that to be a monopoly prices need to have gone up — so Amazon and Google have avoided it
- Famous Manne Economics Institute for Federal Judges part of an effort to educate economics policy
- From 1980 to today, market concentration of largest firms increased in more than three quarters of U.S. industries
- The 1993 book Reengineering The Corporation saw a world that included (tho not limited to) digital automation: “Much of the old, routine work is eliminated or automated. If the old model was simple tasks for simple people, the new one is complex jobs for smart people, which raises the bar for entry into the workforce. Few simple, routine unskilled jobs are to be found in a re-engineered environment.”
- Industrial robots (references the origin of the word robot) as automation too In us to reduce jobs but to empower them in Germany
- Lee Felswnstein’s critique of IBM and other big tech companies for thinking software was “design by geniuses for use by idiots”
- Bill Gates early quote “Show me a problem and I’ll look for technology to fix it”
- A transition from hacker ethic and the riches of corporate digital utopia
- Authors reference the Solow computer age productivity quote from 1987 but implies we never got the productivity gains we were promised (I’ve seen that quote interpreted the opposite way)
- Hal Varian Google chief economist says we mis-measure all the gains new digital models and technologies offer us (ie iPhones do so many things)
- Tyler Cowen and Robert Gordon (his 2016 book) reference for saying we have run out of big problems to solve (And so our productivity gains are reversing to the mean, after the the exceptional “long century” (as Bradford DeLong wrote)
- Tesla over automated and learned the Toyota lesson that as Musk put it “humans are underrated”
- Robot-coiner playwright Karel ?apek wrote: “Only years of practice will teach you the mysteries and bold certainty of a real gardener, who treads at random, yet tramples on nothing.”
- R&D over focused on digital technology which benefited only a few very rich winners
- Authors critical of April 2021 Economist report on future of work and McKinsey WEF 2022 report (on 4IR fourth Industrial Revolution)
- “Machine usefulness”: This is what the authors argue should be our focus because we can control what and how we create
- Computers follow same logic as Jacquard machine loom
- In 2016, influential computer scientist Geoffrey Hinton said: “People should stop training radiologists now. It’s just completely obvious that within five years, deep learning is going to do better than radiologists because it’s going to have a lot more experience. It might be 10 years, but we got plenty of radiologists already.” (Reminds me of Nvidia CEO Jensen Huang expecting fewer will need to learn computer programming)
- Concept of “rent shifting” by business owners to take away bargaining power from workers
- Norbert Wiener of MIT articulated a different path for what could have been for AI
- Wiener: the best model of a cat is a cat (not a robot approximating a cat)
- Deterministic fallacy (of technology) what happened didn’t have to happen
- JCR Licklider followed Wiener with “man/computer symbiosis”
- 4 tenets of machine usefulness: machines should improve worker productivity; creation of new tasks; accurate information (like the web) and better platforms a d markets
- Norman Bourlag saved a billion lives: what would today’s version of the be?
- Chinese surveillance state is the opposite of what we thought the internet and social media would give us (democracy)
- “The pernicious effects of digital technologies and AI on politics and social discourse were not inevitable and resulted from the specific way in which of these technologies were developed.”
- (((I get annoyed that such a big book has so few real examples of their point, for example on page 353, “ digital technologies, which are almost by their nature, highly general purpose, could’ve been used to further machine usefulness – for example, by creating new worker tasks or new platforms that multiplied human capabilities.” But that’s not actual an example, they keep calling for those outcomes but don’t leave me clearer on how to get there
- Rohingya Facebook disaster as example of the AI paperclip parable: the Facebook “stickers” for people to flag a post as dangerous actually registers it as getting more activity and boosted it further
- Jurgen Habermas: “public sphere”
- Research on Father Coughlin’s 1930s anti FDR radio broadcasts shows it limited FDR voters and increased Nazi pro groups; FDR used broadcasting permits to reduce Father Coughlin’s reach (but I wonder how much it’s also about residents not wanting it?)
- Wikipedia as model vs Facebook; Netflix and the evolution of Youtube
- Hannah Arendt “If everybody always lies to you, the consequence is not that you believe the lies, but rather that nobody believes anything any longer.”
- Former US Senator Mark Hanna (1837-1904): “Two things are important in politics. The first is money, and I can’t remember the second.”
- Ida Tarbell, Upton Sinclair and other muckrakers informed the Progressive Era (1897 – 1920) that followed the Gilded Age (1870-1900)
- Automation technology tends to accrue value added to capital (like robotics), whereas those that add tasks give value to workers (like better personalized information for health care, and VR support for shop floor)
- Tools for climate (can it translate to other technologies): carbon tax, subsidized research and regulation
- Break up big tech, the authors argue
- Tax reform: taxing labor at higher rates (at an average of 25% over the last four decades between payroll and federal income tax) than equipment and software capital (more like 5% right now) means automation is encouraged
- Eliminate payroll taxes: don’t make it more expensive for people to work; increase taxes on capital
- Redirecting technology policy is different than industrial policy that picks winners. Like green tech, don’t pick the company or the type between solar or wind but built structure (carbon tax, etc) that encourages wind and solar
- Data collection and user privacy seems like an important role for government oversight
- Jaron Lanier’s data ownership rights
- Repeal 230 (they do think publishers should only be criticized for “promoting” problematic posts, rather than those posts existing)