Did you read the article? Twitter was arguing that the subpoena was incorrectly issued in the first place, not they are somehow above the law.

The subpoena was based on rules that assume that the twitter user's email address and direct messages (like public tweets) are already public. This is surprising since I think most Twitter users would agree that those things aren't expected to be public.

No. I'm pretty sure the subpoena was for _public_ tweets and personal identifiable information (ip address, email, ...).

I think you are confusing personal identifiable information with direct messages. The direct messages _do_ have an expectation of privacy around them, but that information was never requested.

It was for "all user information" whatever that means.

https://www.eff.org/sites/default/files/destructuremal%20Sub...