As a whole, this means there's less energy available to irradiate the plants, but there's more energy in the wavelengths the plants can respond to, so it's more efficient.
I mean... think about it. Every time you look at a tree and see green, that is light that the plant has wasted. Hydroponic lighting is typically that deep purple, red, and blue, that plants can absorb. The goal is to make the plant leaves black, indicating that all light is being absorbed. Otherwise, you're just wasting light.
Everything you said is right. It’s still more efficient to use the suns energy and then distribute the leds than to directly try to use the suns light. Modern hydroponic grow led controllers can adjust the wavelengths as the plants grow
More information here: http://www.webexhibits.org/causesofcolor/7A.html
Hydroponic lighting is optimized for those particular wavelengths in the ~450 nm and ~650 nm peeks. Some plants absorb additional wavelengths, and could likely be targeted with other lights. The key though is that chlorophyll are like mitochondria -- highly preserved between species -- and they evolved to only absorb particular frequencies.
I think this is the main message. People in this thread are talking in absolutes about 'agriculture', but the reality is that there are many, many things that we don't understand well enough at the above-micro-level scale to be able to say 'you need to measure x, y and z and then we can calculate/model the optimal setup to grow something'. People are basically relying on (informed) trial and error, with feedback cycles measures in months or years, so it just takes a long long time for progress to be made - and even then, we usually don't know exactly why something works in a specific case. It's not like 'oh let's run an A/B test for the best color of our 'buy' button for a few days'.