On Futureproof

I picked this up expecting a self-help career book — I’ve been reading a lot of those — and was pleased to find it was something else entirely. Kevin Roose’s Futureproof critically examines both techno-utopianism and techno-nihilism and concludes that neither fits the moment; he calls himself a “suboptimist,” which was a welcome departure from most of what I’ve read lately.

He opens on the techno-utopian claims, most of which reduce to: people have always poo-pooed new technology, those people were Luddites, and they were wrong — look at everything it gave us. Like Roose, I’ve never been convinced. Yes, we got an industrial civilization, and I wouldn’t want to eat as my European forebears did, on tasteless porridge and onions while risking death from preventable disease and breaking my body with labor. But the answer isn’t technology as a blunt solution either: the dislocations are often enormous, and there’s no guarantee that whatever comes next will employ the people displaced by the last wave — especially with predictions that up to seventy percent of jobs may be automatable within the decade.

To meet this, Roose offers a set of rules and, more interestingly, anecdotes. The phrasing of the rules isn’t obvious at first, but it clicks. “Rule #2: Resist Machine Drift” means we have to stop acting like John Henry — we won’t be able to hustle our way into new jobs when the AI can out-hustle us — and instead fight the algorithmic flattening of culture and hold on to our own identity. “Rule #4: Leave Handprints” makes its strongest case through a mid-century Toyota worker who developed an art of spotting weaknesses in machine production, learning exactly when the machines would and wouldn’t work; that matters, because as executives automate the means of production there will be serious failures of quality control, and humans seem better at recognizing human needs on that front. The one I found most interesting was “Rule #5: Don’t Be an Endpoint.” Roose points out that so many jobs now consist of sitting between two technological systems — APIs, in effect — that don’t or can’t talk to each other: a physician’s assistant who reads measurements off one screen and types them into another is an endpoint. I’d never thought of it that way, and I’m glad he put it in front of me.

There are self-help elements, of course — it’s subtitled “9 Rules for Humans” — but more than anything the book is a distillation of good technology journalism with genuinely actionable points for weathering the storm. I came in expecting something surface-level and learned a great deal.