Regarding the Claude block wave, I did a little research and checked the source of all these posts about Chinese finds. So, one X author posted a "leaked" rule that Anthropic bans Claude's accounts. According to him, the account has a hidden rating and is prohibited below the threshold. Deducted points for frequent changes of IP, virtual card, Linux, immediate purchase of maximum and exhausting the weekly limit, granted for the mobile application, including cloud memory and account age. A list of 18 points has already been circulated online, but the author provides no verifiable evidence. It's particularly offensive for Linux here, although you can understand the logic. What is confirmed by Anthropic documents: On October 8, the company updated the usage policy, the new version will take effect on November 12. What was "if we know you broke the rules" became "if we suspect you may have broken them". In the wording of the rules for the prohibition of abuse through multiple accounts, they added their creation in order to bypass the discovery. Also, new points have appeared: you cannot bypass registration and login, pay with unauthorized methods and give access to Claude without permission, including through consumer subscription proxies. Regional policy prohibits using Cloud while physically residing in an unsupported country; Russia and China are not on the list. In the ban certificate, creating an account from such a country is listed among the reasons for blocking. Card issuing country and payment address must match and support. Registration requires a number from a supported country; VoIP numbers, Google Voice and apps are not accepted. In an Aug. 11 document, Anthropic describes authentication using Persona: They might ask for a passport or other document with a photo and a selfie. I don't know what actually affects the appearance of suspicion among anthropes, and I doubt if anyone does. Perhaps the anthropists themselves do not fully know what they have programmed there.