This is a slippery slope - when can you actually "identify an infringing file?" Can I identify an infringing file because I match the filename against the names of popular movies? Probably not. Can my startup recognize infringing files because it calculates the checksum of each block of each file for use in deduplication? Sort of, with a lot of technical build-out, and never with any certainty...

If I see "OMG WAREZ.rar" in a news feed on my homepage, should I remove the file itself, all access to the file, or just the link to the file? What if I have a strict do-not-delete policy for user data? What if I have a strict do-not-download policy for user data (even public user data), to prevent me from accidentally possessing child porn on my work PC? Do I just make an arbitrary and capricious blacklist of file names?

It's absolutely not a valid point against them - their system for "identifying infringing files" was more like a system for "identifying bad-looking file names," which is tangentially related at best.