> And you would cause untold destruction to amazon rainforests, among other things.

Why would putting all the panels in a desert do that to a rainforest on a different continent?

And how much of the Sahara do you think we’d need to cover in PV if for some strange reason we didn’t put any on rooftops, brownfield land, parking lots, car bodies, and (for most of the world) closer deserts?

http://www.wolframalpha.com/input/?i=area%20sahara%20%2F%20%...

Because, surprisingly, various ecosystems including Amazon rainforests are dependant on sand picked by wind from Sahara. Ecology can be weirder than you think, and having intercontinental dependencies is surprisingly common.

Sensitive enough that 0.5% land use change would do that?

I don't remember the exact numbers from studies, but EU-funded research in that area suggested vastly higher numbers as part of removing fossil fuel dependency

I suggest following the Wolfram Alpha link I gave above. My assumptions, to explain the magic numbers visible in that formula, are 20% efficient and 25% capacity factor.

(And the other person was actually saying electricity, which is what WA is inferring in that calculation)