Using fertilizer and labor sounds better than mining and refining stuff like rare earths to me. You could probably even process waste from sewage into fertilizer creating a somewhat closed system this way. On the balance, this seems like a preferable trade off all around if you're going to need to produce light somehow. There's also no reason why this would need to be a monoculture, and they actually talk about engineering a variety of plants in the article.

I don't really see how you can make the assertion that these are two separate problems that can't be elegantly solved in this way. We'll only know whether that's the case once it's actually tried.