@Robert re demand following physics
OK, let's for a moment assume hypothetically there are no physics issues in following the demand cycle. Whether or not your theory is correct, it has to be hypothetical because no one's really done it yet, certainly not for long enough to see if it does or doesn't affect lifetimes.
So now let's look at the economics as well.
Over a given number of years, a reactor constantly maxed out will generate a given amount of electricity.
Over the same number of years a reactor cycling up and down following the demand curve on a daily basis will generate significantly less electricity overall, but the costs (operating costs and finance costs) will be much the same as they were for the maxed-out one (the fuel costs are negligible).
The electricity it does generate (at peak times only) will therefore be substantially more expensive per "unit" than the maxed out one, and the lifetime income will therefore be less, and there will therefore be less profit in it (if it's profitable at all). Does that matter?
If it is significantly more expensive and if it does matter, where does it leave the future of demand following for nukes?
Unmanageable wind power has trouble matching current-magnitude daily demand cycles without storage and interconnects and...
Inflexible nuclear power has **the same** trouble matching matching current-magnitude daily demand cycles without storage and interconnects and...
Today's grid survives an unplanned loss of a few GW of nuke for weeks on end, and will have to be built to do so tomorrow. Tomorrow's grid will therefore also be able to survive losing a few GW of wind for a few days on the occasions when we're becalmed.
See what I'm getting at?
And all that's before we even think about the time it takes to build a nuclear station (not to mention the waiting time before construction actually starts).