Another idea I recently had was extending Karakeep (https://github.com/karakeep-app/), which has robust archiving capabilities today, to enable announcing archived URLs via a DHT and serving the archive bundle as a torrent (similar to how the Internet Archive creates a torrent file for each item). With some work, it would be the equivalent of a distributed Wayback Machine, with each instance able to act as both scraper and server. With a browser extension that could query this DHT, you could retrieve an archived version of any URL that has been archived by cherry picking from the swarm (each URL becomes a magnet torrent URL, very similar to IPFS, but with less bootstrapping required imho).
Distributed web adjacent (https://getdweb.net/).