The AIndustrial Revolution

I keep seeing pieces (from VC firms, consultancies, the White House, etc.) likening the AI rollout to the Industrial Revolution. I'll admit there are correlations (or at least similarities) between the two. But the similarities I see are not the ones being packaged into neatly wrapped sales pitches about productivity gains and shared prosperity. In these pieces, AI and automation often get compared to the assembly line, framed as the driving force behind the rapid growth of manufacturing during industrialization in the US. I don't mean to pick on any single author here, and I don't claim to have all the answers (like LLMs, humans can hallucinate and get things wrong, too!). But what I think most of these pieces miss is the surrounding context of social and geopolitical factors that came into play during this period. We hear about Henry Ford's assembly line, but the fluff pieces leave out the mass immigration and deskilling of work that let factory owners exploit cheap labor for their own gain. They also leave out where Ford got the idea in the first place, which was the meatpacking disassembly lines in Chicago that Upton Sinclair documented in The Jungle (a fact Ford acknowledged in his autobiography). Electricity, as one of those pieces notes, did indeed replace steam engines. And this allowed factories to be redesigned around unit drive (one motor per machine), which made the assembly line possible. Around the same time, Scientific Management (Taylorism) started to take hold, breaking jobs into small, measurable, standardized tasks. Just as the parts on an assembly line became interchangeable, so did the workers running them. Those workers were mostly immigrant laborers drawn from many countries and many languages, who had no real choice but to survive the conditions they were given. Factory owners regularly used those language differences to keep workers from organizing across ethnic lines, but workers pushed back across those divisions anyway. The 1912 Bread and Roses strike in Lawrence, Massachusetts is one of the better-known examples. Strikers won a 15% pay raise, double pay for overtime, and a promise of no retaliation, and the gains spread to roughly 275,000 textile workers across New England within weeks. The same pattern played out at a bigger scale through the 1919 steel strike, the founding of the CIO, and the UAW's 1930s organizing drives, each a direct reaction to deskilling and the punishing pace of the new factory system. Meanwhile, the supply of cheap labor kept shifting. The 1924 Immigration Act eventually closed the door on European immigration, but capital found a new pool to draw from. African Americans moving from the rural South into Northern factories during the Great Migration) filled many of the same roles, and faced many of the same exploitative conditions. Geopolitics played a role too. World War I pushed standardization across manufacturing, largely through the War Industries Board. Through the middle of the century, while Europe was rebuilding, the US (which hadn't been a site of direct conflict) was able to ramp up production on a scale nobody else could match. And coming out of World War II, the Second Red Scare was used to discipline labor in the US, which shaped how much of the new productivity gain workers could actually capture and how much went to the bosses. If we look at the potential AI bubble we're in now, there are also parallels to the Great Depression. Businesses that survived the Depression did so by cutting costs, and the crisis itself forced rapid adoption of practices that squeezed workers (sound familiar?). The companies that came out of the 1930s as winners often did so by extracting more from fewer workers, not by inventing new things. The current playbook in tech looks pretty similar. We see layoffs framed as AI-driven efficiency, more work piled onto fewer engineers, and a handful of "hyperscalers" absorbing a larger share of the industry. AI optimists like the comparison to electrification, and they often point to Paul David's argument that productivity gains from a new general-purpose technology can take decades to show up because factories and firms have to be restructured around it. That's a fair point as far as it goes. But the restructuring David describes is the same restructuring I've been describing in this post (deskilling, Taylorism, workers treated as interchangeable parts). The productivity gains came, but they didn't come for everyone. So yes, electrification played a role in the Industrial Revolution, as did the assembly line. But without looking at the wider context, we risk mistaking correlation for causation. The bigger system around those technologies included a labor regime built on immigration and deskilling, a geopolitical setup where American industry had no serious competition, and a financial environment that took twenty years and a depression to sort out. The technology mattered. But it landed in unusually favorable conditions. Whatever benefits working people gained were won, not granted, through fights for fair pay, safety, and dignity in the workplace.