20M seems like a low number and I’m guessing they all used citations or similar content somewhere on the back-end that would map to NYTimes content as a result of a legal discovery request.
Also down to 20M from 120M per court order.
Sorry, but this seems a completely reasonable standard for discovery to me given the total lack of privacy on the platform - especially for free users.
Also sorry it probably means you’re going to owe a lot of money to the Times.