The point about archiving is actually absolutely correct.

What we do need is one or more archiving relays. Luckily, fetching data from Nostr is much easier than fetching it from mainstream platforms.

I think it could be an interesting project for the Internet Archive or the Archive Team.

It probably wouldn't be too costly, either. It would be acceptable for it to be rather slow and also to have a high latency between when a post is published the first time and when it's available from the archive.