If I understand what you're proposing you want to store the entire articles themselves in the blockchain. That indeed offers many benefits but it's also prohibitively expensive. In practice people will only store small hashes in the tree which can be then used to authenticate offline data. That means that you need somebody to archive that offline data to really make sense of the signed data so you have exactly the issues that you're describing if nobody cares to do it.

The fact that storing data on the blockchain is expensive is also not an optimization problem that can be solved easily with incremental optimizations, it stems from the fact that the data is transferred and mirrored forever on thousands of nodes. If storing data on the blockchain was cheap you'd quickly run into scalability problem. People who run Blockchain nodes don't want to mirror Wikipedia, the output of all the world's newspaper and child porn archives on their servers.

No, of course I'm not proposing anything like that. On the contrary, I was saying that it seemed impossible. Can't you see we actually agree about that!

I pointed out other possible problems with the original suggestion that news agencies should use blockchains. But you seem to completely misunderstand those words as being an attempt to support that original suggestion. Why?

I'm amazed that your reply to me repeats almost exactly what I had just said: "[blockchain could] store a hash of published revisions ... [but] copies of the original articles would be much too expensive to store on today's typical blockchains."

I hope this won't sound rude, but I've got to suggest that you read what people are actually saying, instead of assuming you know what they're going to say.

No wonder you have so many "exhausting arguments".... you're not listening to anyone but yourself.