From the article:
From my understanding, current Solar panels range anywhere between 15-22% in efficency, so at 1%, they are at 1/15th to 1/22nd of what normal panels can achieve, with the 5% figure improving that to a factor of 3-4. Using numbers from
this very good piece on Solar efficiency, at 1% we're looking at around 0.05 kWh/day per square meter for a typical US location, maybe a bit less due to windows usually not being angled properly. That's not a lot, at 5% we're looking at 0.25 kWh/day per square meter, which sounds more useable but still low.
The kicker however is this: The amount of PV cells needed for this is actually fairly small, as it's just the small strips at the sides who are electrically active... the clear area merely diverts the non visible light to the sides of the window. So there is a realistic chance that this might actually be feasible, because even if the energy harvest is small, buildings are huge, and have a lot of windows... the useable surface area would be immense. (though only one/two sides of a building would work well, depending on facing.) It all depends on how expensive the material that makes up the bulk of this technology is, how difficult running the wires would be, how long the stuff lasts/how well you can clean it etc. Something to consider that with less energy delivered into the building by the sunlight, heating cost might rise slightly. Also the compability of these and widows to be opened is to be explored, though from what I understand, "cracking open a window" is not an option in huge 40-story office buildings.
Personally, I think the applications for smartphones etc. are limited, since you don't leave a phone out in the open enough to warrant a fancy screen like this...