What has been rejected is Keynes conclusion that it is possible for there to be a macroeconomic equilibrium including substantial, genuine, unemployment, in favor of a neoclassical "long run tendency" to full-employment.

Accepting that long run tendency in "New Keynesian" economics follows from building the theory on the foundation of the long since falsified neoclassical utilitarian theory of Microeconomic behavior. Samuelson's project to create a hybrid of Keynes' theory and neoclassical micro ~ the project that Joan Robinson famously labeled "Bastard Keynesian" economics ~ didn't work, after all, and so the New Keynesian project is to create space for a simulcra of a Keynesian system in the neoclassical short run out of a set of impediments to the rapid achievement of the neoclassical long-run equilibrium.

Now, a neoclassical long-run equilibrium is the equilibrium which occurs if we project all knowable aspects of today into the future and allow market forces to fully work themselves out. No sensible person with real world experience would imagine that this projection will bear any resemblance to what will actually happen, nor that the long-term experience of economic history corresponds to some approximation of the long run projections of the economic state of previous periods, since crucial decisions will be made in the future that will dominate the "long run tendency" of today's economic state.

The long run could only play out if, in contradiction to all historical experience, we stop making crucial decisions and both economic institutions and technology stop evolving. Hence, as per Keynes witty quip, "in the long run, we are all dead" ~ which is to say, the only certainty is mortality, and also to say, in an economy with living people in it, the people themselves keep erasing old long runs with their actions.

I've been accused of being a Marxist, yet while Harpo's my favourite, it's Groucho I'm always quoting. Odd, that.