The argument seems compelling, the logic inescapable. As hardware doubles its density every 18-24 months, courtesy of Moore’s Law, and as software eats the world, technology willreplace a broad swathe of jobs outright–from burger-flippers to diagnosticians–and atomize many others from full-time positions into gigs performed by many fungible workers. Tech, in short, will eat jobs.

Oh, it will create new jobs too, obviously. But it seems flagrantly apparent that technology moves faster than society these days, and hence it seems very likely that technology will destroy jobs faster than it creates them. What’s more, those jobs it creates will tend to be in fields that emphasize human creativity–ie “tournament” fields with a few winners and many losers.

All of which would be a good thing–as most jobs are crap jobs–except that our society is not built for a world in which more and more people are unemployed. Not unless we implement something like a basic income.

…That’s the argument, at any rate. It’s one I’ve made repeatedly in this space over the last few years. (Echoing many others, to be clear.) But intellectual honesty compels me to admit: the available evidence does not currently support it at all.

If the USA is the canary in our global coal mine–which seems likely, given its high technology and liberal labor laws–then the workers of the world have little to worry about any time soon. “Robots Seem to Be Improving Productivity, Not Costing Jobs,” reports the Harvard Business Review. Total nonfarm payroll employment is far above where it was ten years ago:

Even the age-adjusted employment-to-population ratio has, crucially, recovered almost two-thirds of its Great Recession losses. The trend is obvious. Yes, the tech-eating-jobs argument still seems to hold logical water. Yes, this may be a sharp cyclical rise masking a gradual structural decline. But right now the evidence indicates that “tech is eating jobs!” is vaporware at best. Opinions are interesting, but evidence is what matters.

This evidence is arguably a bad thing–no, really–because, again, tech eating jobs is anoptimistic future … assuming we figure out a new tech-driven socioeconomic structure that shares machine-generated wealth in a decentralized way that still incentivizes productivity and creativity. But it’s also a good thing, because, realistically, as a society, we’re not very good at that sort of restructuring. (Cf my favorite Winston Churchill quote.)

Really we’re not even good at understanding the problem. Consider Derek Thompson’s long, puzzling piece “A World Without Work” in The Atlantic. He doesn’t seem to get that the point of a basic income is to supplement jobs and tide people over through periods of unemployment, not replace work entirely; and, as Mike Konczal points out, Thompson doesn’t really address either the wrenching transition to such a world, or the vicious inequality that it might feature–arguably the topic’s two most important problems.

It seems that we face one of five futures:

1. Tech eats many/most jobs entirely. Could be Star Trek, if we reshape our society to fit. Could also very easily be a world of a small, very wealthy rich minority; a barely larger middle class; and a vast impoverished underclass precariat. (Americans tend to think, wrongly, that such a society would be unstable and soon overthrown by revolution. In fact much/most of the world is already structured this way and has been for many years.)
2. Tech atomizes jobs into gigs, and/or creates new tournament-style Extremistan jobs. Not especially phenomonologically distinct from option 1.
3. Tech creates great, or at least better, new jobs for everyone. This would be pretty good! Not as idyllic as a future where work is completely optional as long as you accept a low (but survivable) standard of living, but pretty good. A minority of people in today’s world have enjoyable jobs that both challenge and reward them. (I’m one of them.) People who are high-profile, or have high-profile soapboxes, are more likely to be among this cohort … so, I suspect, they think it’s a more likely future than it actually is. But it’s at least plausible.
4. Global catastrophe, and/or the Singularity, or something else that makes all this irrelevant.
5. Something much, much weirder.

I actually think Option 5, the “unknown unknown,” is quite likely–but it’s a moot point. The available evidence, to my surprise, is currently pointing towards Option 3. Which is not what I predicted, and is no bad thing at all.

But in ten years’ time? Or even five years’ time?

No one can say for sure, but it seems to me that we collectively face a variant of Pascal’s Wager. It seems to me that it would do us a lot of good, and no harm at all, to at least prepare for the possibility of Option 1 and/or Option 2 — especially now that we have some breathing room.

That’s why I’m watching the increasing experimentation with basic incomes around the world with great interest. True, if jobs keep being created faster than technology destroys them, we may not need that at all. But let’s not put all our eggs in that basket just yet. Moore’s Law, and human ingenuity, are relentless and implacable forces.