If we had them now, nuclear plants would be vastly superior to coal plants for baseload, at least in terms of emissions.
The question is, if a nuclear plant takes 20-30 years from project start to being online, how will it compare with what's available then and over the design lifetime (50ish years). Of course, we can't know for sure, but if storage becomes good enough in the next 30 or so years, the lifetime economics of a nuclear plant get pretty bad.
That said, maybe the economic risk isn't as important as the emissions risk. If renewable + storage doesn't work, and we are still running coal plants in 30 years, that's probably worse than if renewable + storage works and ratepayers paid for an underutilized nuclear plant.
The Vogtle plant reactor 3 and 4 project started in 2006, construction started in 2013, and planned operation is 2021; but these things are often delayed. And that was at an existing site. 20 years may be pessimistic, but not unduly so, based on recent experience.
Unfortunately, people don't believe in climate change enough to acknowledge that nuclear is worth the risk (a risk that would be much reduced by allowing more modern plants to be built).