I'm surprised that they use the PV power for heating coils instead of using heat pumps. I'm sure they've run the numbers and considered it too expensive or too much maintenance hassle.

However, if that's the case you would think that you can cut out the PV step as well and use direct heat from the sun to heat the dirt, by running water hoses though the dirt and through solar water heaters. Should be cheaper and more efficient than the sun -> PV -> heat coils cycle.

Read the article. He is talking about 600C heat. There aren't any heat pumps that

a) give you significant gain when you already have ΔT that large (because 900K / 300K is a big ratio)

b) have a working fluid that avoids decomposition at 900K

Pumping heat from 300K to 900K is not a big gain over heating -- the entire thing is premised on using extremely cheap intermittent electricity during the summer, and your savings are capped at 30%.

I just read Project Hail Mary which suggests a use of solar thermal energy and I thought "ha too bad solar PV is so cheap in the real world this would almost certainly never happen" and yet, solar thermal energy plus heat pumps does seem extremely simple and cheap for this company

Does this become much trickier because they're trying to heat to 600c?

Solar water heaters are expensive, complicated, fragile, and generally immobile.

Solar panels and heating elements are cheap, simple and easily replaceable.

In addition to being at 600 C, this scheme also crucially revolves around keeping capex as low as possible. Resistive heaters have lower capex than heat pumps. In a system with low numbers of charge/discharge cycles the efficiency drops in importance compared to capex.