Arguing with a Techno-Optimist

Andy McAfee and I have been friends for a long time, which means we’ve had years of practice disagreeing with each other in productive ways. Our recent conversation on the Futurology podcast was no exception. We covered AI and jobs, wealth concentration, dematerialization, and the planetary future, and while we found plenty of common ground, the divergences were where things got interesting.

We started with the jobs question. Andy was admirably candid about having gotten this wrong before. Back in his 2013 TED talk, he worried that machine learning would shrink the bundle of tasks humans were needed for, producing mass technological unemployment. The facts refused to cooperate: the rich world spent the following decade in historically tight labor markets. Geoff Hinton’s infamous 2016 advice to stop training radiologists has aged about as well as Hal Varian’s prediction that truck drivers would soon be obsolete. (Varian suggested they could become masseuses, to which danah boyd, sitting next to me at that talk, responded by asking during the Q&A, “Have you ever met a truck driver, Hal?”)

We agreed that the lump of labor fallacy is indeed a fallacy. Where I pushed back was on what the aggregate picture conceals. The fallacy may be false for the economy as a whole while remaining brutally true for individual workers and whole communities. Thatcher’s coal miners didn’t smoothly retrain as something else; a generation got lost. My worry about the AI transition isn’t the total number of jobs. It’s whether people (especially older workers) whose jobs are displaced or radically transformed will be kept whole. To his credit, Andy largely conceded this point, noting that economics carries a blithe assumption about frictionless retraining and relocation that real life keeps falsifying.

The sharper disagreement came over wealth concentration. I pointed out that if Elon Musk becomes a trillionaire (as is likely to happen tomorrow with the SpaceX IPO), his wealth will equal roughly 3% of US GDP, twice John D. Rockefeller’s share at his peak. For someone with small-r republican commitments, that degree of concentrated wealth and power is alarming, particularly given how undemocratic the impulses of some of these men are. Andy’s counterargument was Sweden and Switzerland: societies with high wealth inequality that remain pleasant, egalitarian places to live. He worries far more about polarization and collapsing institutional trust than about billionaires. I suggested those phenomena are connected, since billionaire money in politics is one of the engines of polarization. He thinks they’re largely distinct problems, and that the fix is getting money out of politics rather than worrying about the existence of the fortunes themselves.

On questions of judgment and taste, we had an instructive exchange. Cribbing off my recent Noema article, I argued that taste, design sense, and judgment will be among the durably valuable human skills. Andy got visibly nervous at the word “judgment,” citing seventy years of research showing that human judgment is overconfident pattern recognition, and that when people overrule algorithms, decision quality usually drops. Fair enough. But I countered that as algorithmic predictions asymptotically approach near-perfect quality, the humans nominally supervising them will grow complacent, producing new categories of accident. He agreed, citing the dismal record of safety drivers in autonomous vehicles.

The deepest divide concerned planetary limits. Andy’s book More from Less tells a remarkable story about dematerialization: rich economies now literally weigh less each year. I don’t dispute the data. But I pressed him on whether relative dematerialization matters if total planetary consumption keeps rising, citing Wackernagel’s ecological footprint framework. Andy dismissed such balance-sheet exercises as hand-wavy, arguing nobody knows where the planet’s limits actually lie. I’m less sanguine. And when he asserted that poor countries will follow the environmental Kuznets curve traced by rich ones, I reminded him that we heard identical confidence in the 1990s about rich countries inevitably becoming democracies. That prediction was well-founded in the data of its moment too, right up until it wasn’t.

Andy will take any properly structured bet that civilization in 2100 will be better than today. I hope he wins it. But hoping the curves bend in time is a strategy, and the gap between a two-degree world and a three-and-a-half-degree world is the gap between two radically different futures. It’s all well and good to be a techno-optimist, so long as we continue to stay dystopia-aware!