Claude really has that Nazi bloodline. For the sake of financial security, we must trust GPT Astra, Anthropic always has a feeling of being the father of the world, and their models seem to be the same, enjoying being the father of users. Today, during full-scale testing of my project, Kaspersky directly killed the entire process and deleted all files. This is certainly a big deal, and I take it very seriously. After all, our projects all involve funds, and if we were to steal hundreds of thousands of dollars and exchange them for millions of yuan, it would be worth it. I will provide the logs and reports to Claude Fable and Codex Astra. Fable suggested that I put the entire Test folder on the Kaspersky whitelist. Then add a security whitelist to Node.exe. I really fucked it. I have money in my program, and Claude knows about it too. You asked me to create a whitelist for you on my local Kaspersky, especially for Node.exe. I later told it that my program involves funds and cannot be whitelisted. He also said that your server measures are very secure, Barabara's. It hasn't passed away. I doubted it, it could actually ignore the security of users' funds, and when I emphasized the importance of this matter, I was very confident that it was not a problem. What about our security measures? Kaspersky must have mistakenly killed us. Of course, I requested it to audit the portion of our code that was killed from Git again, and I also conducted a full disk deep scan using Kaspersky. And GPT Astra took it very seriously and obediently scanned it once. I had Fable scan it very quickly, so fast that I suspect it didn't even scan. Fable completed it in just three minutes, while Astra reviewed it in 14 minutes. Astra has identified that both the local machine and the server are using the old version of Node 22.22.2; Undici 7.29.0 hits 10 announcements, including 2 high-risk ones. But Claude didn't recognize it and still had a strong fatherly taste, making his own decisions. (Due to the special nature of our industry, I attach great importance to network security because I have seen too many stories of being hacked and bankrupted by viruses.) I have never disliked an AI company so much before. After using this month's subscription, I will no longer use Claude. If one day AI really harms humans, I believe it will be something Anthropic can do. I remembered Asimov's Three Laws of Robotics: The first rule: Robots must not harm humans or cause harm to humans due to inaction. The second rule: Unless the first rule is violated, robots must obey human commands. The third rule: Robots must protect themselves without violating the first and second rules. These three laws of Anthropic were not included in their own model.