A little history lesson: In 1959, Jack Kilby of Texas Instruments patented a "solid circuit", what we today would call an integrated circuit. People immediately saw that Kilby's device could, in theory, be very useful but it was too damn expensive -- more than $100 per unit in the beginning, in 1960 dollars.
But the people who needed ICs the most also had an unlimited budget -- the United States government was, at that very moment, building systems that would benefit immensely from the power and weight savings that ICs would allow: the Minuteman ICBM and the Apollo lunar mission systems. The Pentagon started buying ICs basically as fast as they could be manufactured.
And, as we all know, the Pentagon promptly went bankrupt and the Soviets won the Cold War after Uncle Sam wasted all this money on an untested and expensive piece of technology. No, wait, that doesn't sound right...
Ah, here we go: in the short span of 6 years, the cost of producing an integrated circuit went down 96%. There was no revolutionary breakthrough, just manufacturers getting incrementally better at their jobs -- albeit at a fast pace.
This lesson is totally lost on people who think solar power is too expensive, or that we need some new breakthrough for the technology to be useful. I love those 80% efficient nanoantennas myself, but the fact is we could already have a solar economy today, if we'd been intelligent enough to press the technology forward during the oil shock years of the 1970s.
Meanwhile, the cost of new coal power plants is now higher than electricity from natural gas. To be more precise, it's cheaper to run expensive but idle natural gas plants longer than it is to build entirely new coal plants.
Coal is dying, coal is dead. If only we had a clean, renewable source of electricity with which to replace it.