Thus, individual small donors with time to do research should expect to be able to find highly cost-effective giving opportunities that GiveWell missed.

This does not seem at all plausible to me. If your research is good, shouldn't you contribute it to the commons? And if you can harness the commons, you should care about the room for funding. GiveWell does what it does for a reason and if you compete with it, you'll become similar. If there is opportunity for more research, shouldn't the answer be for GiveWell to expand or be duplicated? And you do give examples where the answer seems to be to duplicate it, but shift its methods and goals. But none of them seem to have to do with the size of the donor, or the ratio of research to funding.

If there is a proliferation of researchers, they won't have track-records to harness large funding. But how do the researchers themselves know that their research is better than the alternative researchers?

Later you reveal that you were talking about is local improvements. Why not say that up front if that is what you mean? And I think that the relevant difference is not in the quote, but local knowledge that you are unable to communicate to another evaluator.