10 comments

  • tartoran 5 hours ago ago

    There's another discussion thread:

    https://news.ycombinator.com/item?id=48734373

  • kyokoL 6 hours ago ago

    Anthropic has embedded hidden spyware-like code in Claude Code that covertly targets Chinese users. It then sends information regarding every user by injecting it into their prompt message.

    Claude Code is sending info like timezone, proxy and possible AI Lab connections into the system prompt in ways Chinese users can't notice.

    A coding agent with repo and command permissions should not silently hide routing metadata inside prompts. This is a serious breach of user trust.

    • chomp 5 hours ago ago

      I don’t understand why this is a problem.

      • iAMkenough 2 hours ago ago

        Hiding it from users and then charging for additional tokens (if you’re Chinese).

        Imagine the outrage if they did the same to Jewish people.

      • chews 5 hours ago ago

        you are the perfect customer.

    • StopTencent 4 hours ago ago

      [flagged]

      • jml78 3 hours ago ago

        Let me guess, you are 100% with the fact Anthropic trained all their models on our data. Trained it on copyrighted data.

        At least the Chinese companies are paying Anthropic for the privilege of distilling their models.

      • Zambyte 4 hours ago ago

        Copying data without consent is only cool when Americans do it

        • theplumber 3 hours ago ago

          When they don’t need to copy it anymore it will become uncool again

  • dredmorbius 3 hours ago ago