I read these numbers in a profoundly different way from the article. Based on these results, it looks feasible to divert the main mass of a potential impactor.

I read these numbers in a profoundly different way from the article. Based on these results, it looks feasible to divert the main mass of a potential impactor. That appears to run the unsurprising risk of ejecta that would themselves impact. But this just means that the first impact decreases the problem measured in energy by (roughly) four orders of magnitude. This is because the ejected masses are much, much smaller than the original mass and because most of the ejected masses won't impact. Nothing prevents applying the same technique to a few selected ejected masses. If the ratios apply again, we can decrease the thread by another four orders of magnitude. This is enough to reduce the KT impactor to a level that will burn in the atmosphere.