“Proof-Of-Work” Proves Not to Work (2004) [pdf]

It isn’t relevant to crypto. This is specifically about spam email mitigation. I don’t agree with their conclusion, which is “we can’t reduce spam to the level we’d like without imposing undue burden on legitimate senders so we shouldn’t do it at all”. The “all or nothing” doesn’t make sense. They never consider things like whitelists (legitimate senders need to do zero work), different levels of work required for different senders (eg a sender’s IP address isn’t recognized so dramatically increase the work requirements), etc.

It's only relevant in that Satoshi took the proof of work idea for Bitcoin from a spam prevention system like the OP discusses.

A lot of cryptos try to get around proof of work (proof of stake, for example) because PoW takes a long time and uses a lot of electricity, and this is the paper's main complaint -- but so far it's proved quite secure for Bitcoin. The Bitcoin blockchain has never been successfully hacked.