Perhaps not as original. A reader writes:

Such products have been for sale for about 5 years. Unfortunately, its not as simple as that blog entry contends. With silicon flat panels, you are not so sensitive to the direction of the sun. With a magnifying glass, you need to be pointed directly at the sun or all the light focuses somewhere else. This means that you need mechanical trackers turning the units continuously. This adds maintenance cost, etc. In order to get as much power as possible, they use so-called triple junction cells which are about twice as efficient, but substantially more expensive than silicon cells, and made with much more exotic and less prevalent materials than silicon (second most common element in the earths crust). This doesn't mean it won't work, because you are tracking the sun with the unit, you get about twice as many hours of effective sunlight which can help. But then you have to space the units further apart which means peak watts per sq ft are lower. Oh, and these cells won't work if it is cloudy, there is nothing to focus on the chip where as silicon cells still generate power even on cloudy days.

So…for places where real estate is cheap, sun is prevalent, and labor costs are low (think Mojave), concentrated solar power may have a chance, but that has yet to be proven. So far, the cost is 2-3 times the cost per kW-hr of silicon. And for residential roof tops, no way. Or for places where the sun isn't a constant (think Seattle). Lastly, the silicon guys have so much head start in terms of volume that it is going to be hard for any other technology to catch up. Its one advantage is efficiency, which if carefully managed, could lead to more watts/acre for a utility type installation. To get more involved, the typical silicon installation only has about 40% of the cost in the cells. The rest is installation, mounting hardware, etc.

Any technology will have this cost, and a reasonable approach is to try to win the race on 'balance of system' costs. Efficiency is a big key here, since wires, mounting pedestals, etc. all scale with the power generated per cell. Silicon manufacturers are getting 21-22% out of their cells now with a promise of ~28% in the near future, this will make it very hard to beat, even with very cheap solar cell materials. If the material is hyper efficient and using concentrators, the cell can be nearly free but if the hardware (trackers, etc) is twice as expensive, you haven't won. Of if you are using some of the new 'thin-film' technologies where cells are inkjet printed, you again get 'free' cells but at half the efficiency (~10%) and twice the number of cells needed and twice the balance of system cost. Still, the scale of the problem is something to consider. To replace current electrical demand in the US with photovoltaic, you would need about 10,000 sq mi of silicon cells. Not to mention the systems to store power for night, and to transmit it to low solar flux areas. And as we convert more to electric from hydrocarbons, the amount of energy needed will only increase.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.