It could also be signed using SSL or GPG or really just any other kind of PKI. In the end I need to establish a root of trust, so I either way I need to know that 2c414...2d7 is indeed the hash of the chain_address/ssl_cert/root_ca/gpg_key/... - so in that regard there is no advantage of using a blockchain over a "traditional" PKI.

Amend: The real question is how to verify cropped and recompressed snippets. In academia or HN we add a "see foo et al. [1]", so probably that needs to be encoded into the data stream, e.g. "authoritative src is https://media.whitehouse.gov/permanent-url/1234[...] with hash $hash signed by $cert_hash on $data, portions are $some_encoding_here" and better be signed as well.

Did not mean to be dismissive of your point. It is just that the overhang of blockchain as a solution to everything is so prevalent that sometimes need intervention.

It is true that the cryptographic primitives that blockchain is built on are relevant to many other things that we can do and have done. Blockchain in fact utilizes those properties to build a distributed shared state on which the parties have consensus. That, however, is a very special use case of those primitives. If you need message authentication, digital signatures have done that for decades and you don't need an expensive associated database to accomplish that.