#95: Stuck in the Middle With You
Recently, after threading my way through the platoon of people who hand out flyers on the busiest part of Bedford Avenue, I realized that gig—pressuring fast-moving strangers into taking a leaflet and then engaging them in some sort of sales pitch—will be the last job that gets automated. Imagine a robot trying to stick a flyer in the hands of passersby, and how much more easily everyone would ignore that robot without any eye contact or cursory acknowledgment. However unskilled this job is, it’s fully human. Keith Johnstone, in Impro, his book about theatrical improvisation, even cited flyer distribution as a prime example of social space dynamics: “You can’t just thrust your hand out at people, you have to establish that you’re giving out leaflets, and then present one at the right moment.” No one ever wants to have a flyer handed to them, yet people do actually take them, and sometimes even stop to talk. That interaction is incredibly dense with interpersonal empathy and status jockeying. A machine, on the other hand, could perfectly time the flyer handoff and still not do the job any better than a table would.
Marshall McLuhan said machines are “extensions of man” that focus and intensify our innate abilities in specific directions. This is more obviously true of machines that do physical labor. Digital automation, on the other hand, is most promising in areas that never came naturally to humans but were themselves enabled by previous technologies, such as operating other machines (driving) or doing math. Even memory, as performed by our brains, works differently than the computer version, which captures everything without selectively prioritizing what’s more meaningful; for those kinds of tasks, computers don’t extend or amplify but simulate, and now that we expect computers to be capable of everything we’ve historically done, we witness more examples of computers simulating human faculties poorly, and butting into culture awkwardly. The internet, particularly social media, is replete with this: Human products like meaning and status, when they appear online, as this essay argues, are frequently residual, “the result of previous work,” and are thus easily manipulated, more than they are “extended.”
You may have seen this glib Farhad Manjoo column a few weeks ago about how “the global economy runs on parties you’re not invited to.” He writes, “Even though we have all given ourselves over to data, human relationships lubricated by human pleasures still matter to every industry.” Despite one vision of the future as a huge, all-encompassing efficient market where everything is deterministic, the humble human still plays a decisive role, freed from the tasks we’re bad at to do what we do best: talk, sell, manipulate, hustle, scam, play golf with one another, and party. The most durable role for humans in the near future that’s actually coming, then, involves a strange role reversal with software: Our job is to function as the interfaces between inscrutable automated systems of various scales whose internal operations proceed without our involvement, but can only extend their reach with our help. Instead of eliminating the middleman, digital platforms have solidified that as our permanent role—the most human job of all.
Reads:
Geoff Manaugh on the “drowned lands” of Upstate New York, former islands that have just become landlocked hills ever since a controversial project to dam and drain the formerly flood-prone region. “A half-century of ‘war’ broke out among local supporters of the dams and their foes,” in which the dam-building faction was called the “beavers” and the dam destroyers were called “muskrats.”
McKinsey as “capital’s willing executioners.” A nuanced analysis of the firm’s powerful yet relatively invisible role in the world, and its structural inability to solve the most urgent problems. “The firm has an enormous stake in things continuing more or less as they are.” Published anonymously by a former employee.