My thoughts on unemployment and the business cycle have been deeply influenced by a brief paper Robert Gordon did on what he calls "1978 Macro," so I'm persistently saddened that instead of following this up with more work on the subject, he's going around doing talks on his thesis that innovation and economic growth are dead.
Fundamentally, it seems to me that even if this were true there's no way Gordon (or anyone else) could know that it's true, so it's an odd thing to be so confident about. But I do have a specific grounds for optimism that I think Gordon misses in his discussion of the limited economic benefits (thus far) of the information technology revolution. This comes from the economics profession's unfortunate habit of rhetorical slippage between English-language words like "technology" and "innovation" and the technical concept of Total Factor Productivity from the Solow Growth Model. Something we see from Europe in the centuries before the industrial revolution is that it's perfectly possible for amazing innovations—the printing press, eyeglasses, calculus—to have immeasurably small effects on productivity. A printing press based on movable type, for example, was an enormous boon to productivity in the book manufacturing sector. It had almost no impact on economy-wide productivity, however, simply because the book manufacturing sector of 17th-century Europe was trivially small. You don't get the Industrial Revolution and rising living standards until the trend toward improved technology starts affecting a sector that's really big and important—apparel manufacturing.
And my guess is that's how we should look at digital technology so far. It's not that the technology itself isn't important or transformational. It's that the sectors it's transformed are, themselves, not that important. We've had a productivity boom in the journalism sector, for example, but the 21st-century journalism sector is like the 17th-century book manufacturing sector—it doesn't matter to the economic big picture.
But if an equivalent transformation had occurred in the much larger health care or education sectors, that would have been a boon to growth and wages. In other words, if future applications of digital technology make it cheaper to get a college degree or a medical diagnosis the way they've already made it cheaper to read the news, that means the real wages of every waitress, truck driver, electrician, interior decorator, and architect in America go up. How likely is that to happen? It's hard to say for sure. Future technology is unpredictable. But having lived through the ongoing transformation of journalism, it looks pretty likely to me. The business of teaching people stuff or informing them about which illnesses cause which symptoms seems to fundamentally have a lot in common with the dissemination of news and news analysis.