Pretty much everything in AS/AD is riding on the hypothesis that labor supply is highly elastic at the nominal wage and labor demand is reasonably elastic at the real wage. There is nothing for entrepreneurs to figure out–they will employ more workers as long as you can trick those workers into taking lower real wages.
That is Arnold Kling, here is more. Many Keynesians like to poke fun at the RBC or rational expectations ideas of unemployed workers taking a "voluntary vacation" during a downturn. Yet Keynesian theory has a no less serious problem, namely that workers take a "stupid voluntary vacation" during the downturn, due to their stubbornness on nominal wage cuts. Reflation, if it comes, is doing no better for them in real terms than if they cried Uncle in the first place and simply lowered their wage demands.
The more dire a story you tell about the costs of unemployment, the more embarrassing the puzzle becomes on the side of positive economics.
Aggregate demand macroeconomics works in many cases and it almost always "works" (predicts well) when the macro forces are pointed toward destructive ends. We are not sure why it works at all, or if it always works, and yet we see a great fervor of belief in it and a demonization of those who are skeptics.