The End of Intelligence

What Happens When Intelligence Matters Less

In the 1950s, Detroit was the richest city in the world. By 2008, we were bankrupt.

There were many factors in Detroit’s decline, but growing up, the most visible was deindustrialization. Driving into the city each morning as a teenager was like passing through a cemetery. Tombstones of ruined factories. Teetering warehouses. Cavities of blight. I joked with my friends that Detroit was the ‘largest rural town in America,’ because large swathes of the city were given back to nature, converted into overgrown fields. Things have gotten a lot better since then; Detroit really is bouncing back after decades of hardship, and the city is unrecognizable ten years later. But with its broken-down buildings and empty fields, Detroit taught me a cautious lesson:

It can happen anywhere. It happened here.

Packard Plant. Photo credit to Camilo J. Vergara and the Library of Congress.

The reason I make the comparison is that I believe we’re on the precipice of another seismic economic transformation which, like deindustrialization, will restructure American life. Detroit had a lot of problems, granted, but deindustrialization kicked its teeth out in the 1970s. More generally, deindustrialization ravaged the American Midwest and undercut the middle class. Similarly, the coming economic transformation will vitiate the professional class, targeting HCOL coastal cities like New York, Boston, and San Francisco. Just like deindustrialization, the deathblow will be the one-two-punch of outsourcing and automation. Knowledge-work in America is coming to an end.

This isn’t a hyperbole, but it will take another ten years (more or less) to become self-evident. Right now, knowledge-workers already feel the creep of outsourcing. Every month, more and more American knowledge-work is shipped off to India. Although companies complain about poor communication and low quality-of-work, these are just growing pains, functions of India’s rapidly expanding economy, and this grumbling isn’t stopping corporations from dumping business into the subcontinent. In a way, this was inevitable; a globalized world meant a globalized labor-force, and Internet technology allowed professionals to work from anywhere. Still though, you have to admire the suicidal shortsightedness of the American professional class. They clambered after work-from-home, either not realizing or not caring that they were proving their replaceability. If you can work effectively as a Digital Nomad in Bali, then I can replace you with someone from Buenos Aires or Bangalore. Geographical proximity is no longer necessary; the office is an outdated concept. American knowledge-workers have sowed the seeds of their own destruction.

But you also have to feel bad for this professional class—I’m one too—because the much larger threat isn’t their fault. Automation is coming. AI technology trims the difference between man and machine, making machines competitors for knowledge-work employment. These machines excel at basic customer service and procedural writing. They’re indefatigable employees. They’re easy to implement. I’m no Marxist, but I do find some useful grist for understanding this coming change in Marx’s descriptions of industrialization. Envisioning the effect of industrialization on wages, Marx notes that, “since the worker has been reduced to a machine, the machine can confront him as a competitor.” Interchangeable employees with no legislative safety nets can easily be automated and replaced. You don’t need to buy Marx’s discussion of alienation (the mechanization of the worker) to recognize that machines are competing with employees, and human beings are losing the competition. Communities are left behind, and more will be soon. Just ask Detroit.

This difficulty compounds with the advent of human machines. AI models increasingly exhibit human capabilities. They’re smarter than many Americans. They’re capable of automating—either fully or partially—everything from wire reporting to patent documentation. These industries are already exploring AI, and the success of their early experiments has triggered an abrupt hiring slow-down in the professional class. Even if you don’t think AI can replace you, your boss does and is open to testing out the theory. This kind of automation doesn’t require the ‘Singularity,’ some pure machine intelligence which is free, conscious, and rational. Indeed, some of these characteristics might even be impediments to automation. In other words, AI doesn’t need to be a human being to automate human knowledge-labor. It just needs to continue improving, and odds are that it will.

In practical terms then, we’re looking at the end of knowledge-work as the dominant form of American productivity. The end of intelligence, in the sense that being smart won’t equate to much economic value anymore. Our current incentive structures reward intelligence—doing well in school, passing recruitment IQ tests, thinking creatively to find new optimized angles in staid industries. But if you’re a regular person, the typical A-student, then you probably can’t out-think the machine, and you certainly can’t outwork it. The machine will subsume knowledge-work jobs which require good writing and a liberal arts education. Thus, companies won’t need ‘smart’ people anymore—college-educated employees who specialize in verbal skills and analysis (law/business/coding, etc.). Look at dreary market drop-offs facing recent CompSci and MBA graduates. These degrees are little more than booby prizes. And even if automation is only partial—as it likely will be—then this still constitutes the decimation of the knowledge-work professional class. A skeleton crew of supervising employees left to grease the chittering machines.

The future looks grim, and twenty years out, American starts to look very different. I don’t tell you this to scare you, and I’m not necessarily convinced of its inevitability. Trailblazing legislation, for instance, could head-off these threats, and there will be political incentive to do so before long (just like deindustrialization became a reliable political trumpet). But let’s think through the effects of automation and outsourcing—deintelligizing—more tangibly.

Just visually, consider the empty office space. The death of knowledge-labor will signal the end of the ‘sinecure’ position for millions of young Americans. No more project managers. Bye-bye MBAs. Six-figure salaries will become far less common, and the road to the upper class will seem less meritocratic and conventional. There will be no safe path to prosperity. This frittering of the well-paid professional class will hollow American cities built around that brand of knowledge-work. Where previously, these coastal metropolises pulled droves of young people from Midwest ruins, now the dynamic will reverse. The combo of outsourced jobs and lowered wages will render cities like New York unlivable and unaffordable. These cities will bleed population as Americans attempt to cut their teeth in less competitive and overloaded service economies. And the sudden vacuum of high-paying professional labor will boost crime, making these places even less appealing to live in. COVID already initiated this coastal exodus, even before deintelligizing got going. This should give some perspective to what I’m saying. I’m not describing some speculative future. I’m laying out, with brutal honesty, the progression of trends which have already begun.

Norbert Wiener makes a great point in The Human Use of Human Beings. Only in the last few hundred years has technological change become truly meaningful. Before, say, the Renaissance, human society looked basically the same on a technological level as it did a thousand years prior. Wiener’s point is that society’s rapid progress isn’t a guarantee. It’s an artifact of a specific configuration of technology which is contingent and mutable. We should hedge against the truisms of our current market system, like the notion that technology creates more jobs. One can easily imagine future scenarios where technology wouldn’t create labor opportunities. I’m not sure AI is that scenario, but consider the writing on the walls. If AI continues to automate intelligence—as has already begun—you’ll be competing with an economic actor much faster, smarter, and more ruthless than yourself. It’ll be like a chimpanzee trying to outsell a human being at the marketplace. At best, the new jobs created by AI won’t be knowledge-work, won’t promise the generous wages of professional labor today.

The automation of intelligence means that the era of the professional class is fast coming to an end.