> I still don't understand the economics when it comes to power all the time, not some of the time

The way a traditional grid works is that you have baseload plants (nuclear, coal) that generate all the time and peaker plants (hydro, natural gas) that make up the difference between the baseload generation and the current demand by varying the amount they generate to match demand in real time.

The higher demand periods when you're not using electricity to heat buildings are typically daytime and early evening. Solar generates power during daytime. That makes "use solar instead of natural gas during daytime" an easy win. You can also do things like "charge electric vehicles mostly during daytime" and use solar again. Then you're still using natural gas in the early evening but you save a lot of fuel (and CO2) by not having to use them during the day. Meanwhile the gas plants are still there to use in the evening and then you can use them on a day when it's cloudy.

That's still where we are in most places. There isn't enough solar in the grid yet to completely replace natural gas during most of the solar generation window, and we could add even more by electrifying transportation, so we can still add a lot more solar before we have to really deal with intermittency at all.

Optimists would then like to extrapolate the economics of doing that to doing 100% of generation from renewables, which would require actually dealing with it.