These are very good points of course.  I did my own maths, based on about 500W for 6 hours a day, which is power I should have pretty much "always", even in winter.  That would be enough to heat all the water we need.  But yes, of course, it would mean that the miners do not run all the time, which is of course a thing to factor in.  That's why I was also thinking about running especially cheap, less efficient old miners - because in essence it would still be better and more fun than just burning the electricity anyways to heat water.  Some people (including "experts") I talked to suggested not getting a thermal solar installation anymore, but just PV and simply doing that instead for hot water.

It is also a good point though that of course it won't be worth it if it just generates $10 a year or whatever.  So that is indeed something to consider, especially as difficulty goes up and rewards are halved.  I will have to carefully look at the hardware suggested here (S9 and S19 with some boards disabled) to see what makes sense for me.

The alternatives would probably be getting a heat pump to more efficiently use the electricity (but then, the heat pump itself is a lot more expensive than a S19 with watercooling setup), or just keep using the oven (burning wooden pellets in my case) and selling the electricity back to the grid.

The discussion is quite interesting though, and I think it shows that the individual details matter very much - and that might be why miners for heating have not yet been a thing used more widely.