There was discussion about implementing Hashcash for Lemmy: https://github.com/LemmyNet/lemmy/issues/3204

A scary possibility with AI malware would be a virus that monitors the internet for news articles about itself and modifies its code based on that. Instead of needing to contact a command and control server for the malware author to change its behavior, each agent could independently and automatically change its strategy to evade security researchers.