The author makes an argument that is somehow both adhom and slippery slope. By pointing out something that everyone knows (Torvalds is rude, which not necessarily a bad thing depending on who you ask) he adds weight to his second argument about /dev/random being insecure (which is complete bollocks). I christen this logical fallacy the "Ducklinism".

And I have to agree with you about his obvious lack of expertise in the department. Where he fails to make a compelling article is not his lack of expertise, rather the lack of interviewing many people who do have that expertise.