I like the part where he "supported" his thesis that a certain class of comments is "spam" by checking a random sample of the associated email addresses against

I like the part where he "supported" his thesis that a certain class of comments is "spam" by checking a random sample of the associated email addresses against HaveIBeenPwned and finding that 75% of them showed up in there, but did not bother to make sure this is actually a real signal by checking a random sample of "real" comments as well.