Despite Donald Trump’s promises to save factory jobs, the next few years will see the maturation of technologies that compete with blue collar labor. Fighting that fact is a fool’s errand.

Consider a few recent developments on the retail and transportation front. Just last week, Amazon announced Amazon Go, a grocery store experience that does away with checkout lines. Yesterday, Google said it had placed its self-driving car business into a dedicated company called Waymo, marking another step toward the commercialization of autonomous vehicles.

Neither check-out free shopping nor self-driving vehicles have achieved commercial scale yet, but when they do, they’ll threaten certain jobs — like cashiers and cab drivers. Resisting technologies like these is a waste of time — and a distraction from the task of adapting our workforce and educational system to the new reality they herald. (Technology creates jobs, too, of course, but there’s evidence this is happening less consistently.)

Kids have gotten used to hearing that the jobs they’ll do as adults haven’t even been invented yet. But are they also used to hearing that their not-yet-invented jobs may also disappear within the course of their lifetimes, rendered obsolete by new technologies? Are they equipped for that kind of change?

Artificial Intelligence and White Collar Jobs

White-collar jobs aren’t necessarily safe, either. In an interview this week, Jack Clark, a researcher at Open AI, alluded to the potential for artificial intelligence to automate not only blue but white collar jobs as well, like insurance actuaries and customer service reps. “We now have a methodology to automate the people in these roles,” he said. Clark didn’t say this callously — he also talked about technology’s potential to exacerbate inequality and the need to address this problem with heavier investment in “lifelong education”.

Notwithstanding recent electoral rhetoric, we can’t, and shouldn’t, roll back the technological tide — our economic competitors certainly won’t. (China, for instance, is investing heavily in AI.) So we have to invest, as Clark says, in “lifelong education”. But what does that even mean?

Last week I attended an after-work event featuring two tech investors. During the Q&A session, someone asked the speakers if they knew of any education startups selling products or services to benefit economically displaced workers.

“Khan Academy is free,” the questioner mused, referring to the online education site, “but what about people who don’t even know about it, or lack the internet access necessary to use it? How can education startups address that market?” (I’m paraphrasing his question a bit, but that was the gist.)

There are so many assumptions in that well-meaning question: That the laid-off factory worker whose old job is never coming back should know what new skills to study (coding? web design? marketing? massage therapy?) in her quest for new employment, or that a site like Khan Academy provides sufficient training to make her employable. Or that she’ll be able to find new work in the town she currently lives in. (If factories are leaving her area, there’s a good chance other businesses are too.) Or that she won’t face stigma, or ageism, or plain old skepticism, when trying to find employment in a new industry. Or that the learning curve she’ll face on a site like Khan Academy won’t be too steep to climb in the months before unemployment benefits run out.

One Technology that Isn’t on the Horizon: Time Machines

So where does this leave us on the education front? Do we simply focus on expanding access to higher education? If so, on whose dime, and in what time frame? Do we make computer science a mandatory subject in primary school? (And at the expense of what other subject?) Do we create an entirely new business model focusing on mid-life education — and what does that look like? Do we (either the private sector of the government) provide better career intelligence and labor market forecasting within schools, colleges, workplaces, and unemployment offices?

I don’t know the answer, but I know this: One technology that won’t be invented anytime soon is the time machine — so there’s no turning back the clock. We must look ahead and innovate our way to a solution.