So sleep reduction drugs like Modafinil push down wages in the short run, but that increases the demand for capital which pushes wages back up in long run. If the economy is straighforwardly scalable—if doubling the machines and work hours exactly doubles output—then the sleep reduction drug will have exactly zero effect on long run hourly wages.

And you don’t even need incentives, optimization, choice, all of that to get this clean result: The original Solow Growth model showed that even if people just consume a fixed fraction of their wages, and don’t raise their savings rate at all when the productivity of capital rises, you’ll still get the clean result that extra work hours have zero effect on long run hourly wages.

The end of sleep is unlikely to push down wages.