The core problem with Scribd is architectural. A single centralized party sits between the content and the people who need it, and it controls access terms indefinitely, even for documents it never created.
Nostr sidesteps this by design. Long-form content is published under NIP-23 as `kind:30023` events: signed, addressable, and replicated across independent relays instead of stored under one party's control. Once an event is published, no single relay can revoke access to it since any other relay can hold and serve an identical copy. This already works well in practice today.
For binary files (PDFs, scans, media), Blossom extends the same model: content-addressed blob storage keyed by SHA-256 hash, mirrorable across independent servers. I still need to study Blossom more to understand its current maturity level, but the principle holds either way. No single entity ends up as the sole gatekeeper of content it doesn't own.
@fiatjaf, is this something worth exploring further to fix the Scribd-style lock-in problem for good?