Claude that may scare many of companies;

Anthropic's Claude AI model utilizes a hidden workspace for internal thought processing. This 'J-Space' allows silent reasoning and planning, separate from user interactions. The AI's internal workings reveal similarities to human conscious thought processes. Anthropic also embeds code to restrict access to its AI models from China. This hidden processing raises concerns about AI alignment and potential misuse.