Ah yes, a paper from Fan Shanhui's group at Stanford, which kicked off the whole business. Should have linked that instead!

This is 5.7 deg (using metasurfaces to selectively reflect reflect absorb different wavelengths)

compared to a less fancy 4.9 deg from their seminal paper (2014)

https://www.nature.com/articles/nature13883

Maybe comparable cooling can be achieved with even cheaper techniques..?

The interesting implication is that PV has more upside growth potential than nuclear, against the ultimate limit of global thermal pollution.