Sure, that's an excellent point. In case you read late replies: we recognize this, and parameterize the burst energy [1] by a beaming scale factor, \Delta\Omega/4\pi. (It still drops off as the inverse square, though, as long as you're not in the near field.)
The problem is, while you can reduce the energy requirement by making the beams narrow, that increases the total number of sources by the same factor. When we say that there are between 5000-10,000 FRBs all over the sky every single day, we are referring to FRBs beamed towards us. If you apply a 10% beam, your energy requirement drops by 10x, sure, but the source count goes up by 10x.
Right now, we don't have a large enough plausible progenitor population, even at 1x, for these FRBs. It's a really fun problem.
The problem is, while you can reduce the energy requirement by making the beams narrow, that increases the total number of sources by the same factor. When we say that there are between 5000-10,000 FRBs all over the sky every single day, we are referring to FRBs beamed towards us. If you apply a 10% beam, your energy requirement drops by 10x, sure, but the source count goes up by 10x.
Right now, we don't have a large enough plausible progenitor population, even at 1x, for these FRBs. It's a really fun problem.
[1] See, e.g, the Methods section in https://www.nature.com/articles/nature20797 (or https://arxiv.org/abs/1701.01098)