I'm not sure I follow your conclusion. Who buys a solar array for free? It seems to me, a solar owner pays quite a bit to generate electricity for both themselves and the grid, they are compensated for power they deliver to the grid - as are all utilities when they finance generation - through the rate base over time.
The cross-subsidization argument is: since the transmission infrastructure payment is embedded in the retail rate, that the solar customer is being subsidized by others when they are credited at the retail rate.
It is true that when the solar customer uses electricity from the grid, they benefit from the entire grid infrastructure. What is missing from that evaluation is the fact that the solar customer's neighbor is getting power from the solar array that doesn't use the transmission infrastructure and is largely confined to a small portion of the distribution grid. That customer pays the full retail rate to the utility for power that the utility didn't generate and didn't distribute.
My point to Hops is that the transaction is essentially like the neighbor buying power from the solar customer, but using the utility as a conduit - the utility is made whole, the customers are made whole - no subsidization.
That doesn't mean there isn't a huge problem for the utility - they are receiving less in revenues. But as the author points out - that is the case with any type of demand side management, including LEDs. But this is a problem of our current utility business model, not of either efficiency or solar.