I'm not advocating shutting down working reactors until their planned end of life, so we already have a built-in phasing out schedule, measured in decades.

There have been a whole bunch of studies that say you could, in principle. If things like power-to-gas or cryogenic storage panned out, going from 'in principle' to 'in practice' would no longer seem particularly far-fetched to me...

The main problem that remains is organizing the industry to do it when natural gas is dirt cheap due to fracking.

Also, wind + solar is generally a pretty stable mix. We already know how to do continent scale power grids to shift power around local weather. What recently happened in Texas was a result of their conscious decision to go it alone with their own grid, and making insufficient winterization investment in the grid they built.