> Would it make a difference if we were talking about articles on a news website.

News articles are pretty explicitly copyrighted and published for a commercial purpose. The websites make their terms clear when you visit. I don't think anyone can argue that it is legal to copy and distribute these articles, same as a book or movie or song.

Data posted on Bluesky on the other hand is meant to be broadly shared using the AT protocol. It is quite literally a feature. If you create your own Bluesky client, for example, you aren't committing copyright violation by downloading someone else's posts on there. Similarly, you aren't going against any terms of service by consuming a firehose of data from an AT relay.

Right, that's why I asked about Bluesky's content license; just because it's not in your face when you visit, doesn't mean you don't have to abide by it.

You understand that categories of usage are important, right? No-one is breaking the GPL by reading source code, but incorporating into your own codebase can be problematic if not done correctly. Similarly, human beings reading the data posted by a Bluesky user is not the same as aggregating and analysing the data of thousands of users. As I said I'm on the fence with this, but I do understand why someone might have a problem with it.