There are many errors in his calculations. I was part of a team that studied Solar Power Satellites when I worked at Boeing, so I think I have better data than Mr. Murphy. Let me list the ones I spotted:
* He quotes the performance gain against one of the best places on Earth (the US Mohave Desert) at 3:1. For the world as a whole, the gain is more like 7:1. Most places have much more clouds and thus less available sunlight.
* He assumes geosynchronous orbit. This is not required if you have a constellation of satellites and electronic steering of the beams (something every cell tower does today). A lower orbit allows using smaller satellites and ground receivers, closer to the cities that will use the power.
* He assumes 100 GHz transmission frequency. Generally microwave amplifiers are more efficient at lower frequencies and less subject to rain fade.
* Calls launching a large dish prohibitively difficult, while ignoring that the International Space Station demonstrated assembly of large space objects from smaller pieces.
* The study I worked on showed that 98% of the mass of a solar power satellite can be supplied from material already in space (asteroids, and the Moon). Therefore launch cost is not a major issue, provided you set up equipment to extract items like silicon and aluminum from rock.
* Space solar arrays are already 15 times lighter than terrestrial panels, because they don't need frames, seals, and glass to survive high wind and weather. If they were made in space, they would be lighter still, because they would not need to survive launch loads or include deployment mechanisms. Large space solar arrays are launched folded up to fit in the rocket.
* Radiation damage is not as severe as he assumes. Actual space solar cells use Ceria-doped cover glass for protection, and function quite well even in high radiation parts of the Van Allen belts.
* The mass delivery ratio of 100:1 he quotes is way off. Given that you are building big space solar arrays, you can attach electric thrusters to put them in position, something that modern comsats already do. Electric thrusters are about ten times more fuel efficient (although slower). With sufficient traffic to orbit, there will be an incentive to use better propulsion, reducing the ground-to-low orbit ratio by a factor of 5-10. As I mentioned before, with in-space production, you only need to launch about 2% of the satellite mass.
In total, his general method of comparing ground to space solar power is reasonable, but he misses important information and the numbers are way way off.