Interesting report by Anthropic which offers Claude AI: Chinese AI companies (seven of them) were routing prompts from users to Claude and relaying back the responses while recording the interaction to train their own models (called distillation). They found the prompts were coming from not just regular Chinese customers, but state security linked entities were passing sensitive data, from surveillance footage in China to a hacked Russian government database. What I found interesting is, why would Anthropic publicly reveal this and alert Chinese authorities that sensitive data from China was coming in? They could have simply identified the relays that Chinese AI companies were using and given them less capable Claude models while silently siphoning more sensitive data while informing US intel. Of course, this also raises questions if Anthropic reads and analyzes all our prompts and conversations. Its public face of providing great privacy and being all American may be a facade. IMO, all AI companies in the US seem infiltrated to an extent by China using their plants, compromised management, and ideologically brainwashed employees. And, I think, these infiltrates are going to be used to stall AI development in the US soon, with strikes, protests, sabotages from the inside, as the gap between US and China widens in AI. Also, American public who are already being psyop'ed against AI via SM and political pressure from the GLISCO affiliated will also be used to stall AI development in the US. Which means if OpenAI has its IPO after midterms, there's a good possibility it will not do well. And all AI companies may suddenly not do well causing the AI related market boom in the US to reverse by next year.