None of these address the inherent problems the technology has and which the aricle brings up, and it also directly contradicts the claim that you "could built them out much larger". You still need receiving arrays on the ground which have to be of comparable size to the solar parks you're aiming to replace.

Arguing generation efficiency is pointless, as the ratio is pretty much fixed and improving efficiency further will impact both space and terrestial arrays. The point is transmission and conversion efficiency, which is where the technology fails, because of the massive losses.

"Networking plants in orbit" will only exacerbate transmission losses and does not change the inherent limitations placed upon receiving arrays.

I don't even know what the last point is supposed to mean. Are you under the impression that we currently share power the same way we "share" internet connections? That Africa's powergrid is connected to the Americas'?

You can not feasibly eliminate any existing infrastructure with this. Ground based grids will always be more efficient than "beaming" anything. Geostationary orbit is 35000 km up.

Putting things "in space" doesn't magically solve problems, regardless of how cool it sounds. In this case it has more downsides and creates more problems than it actually "solves", not to mention making critical infrastructure exemely hard to maintain or replace. If anything it might make more sense to simply reflect sunlight. That way you at least don't end up with a completly useless groundstation if something on your space array breaks down.