I personally feel that the use of AI has gone through two stages Stage 1: Freedom of Information All publicly available knowledge in human history has been distilled and compressed by LLM. I can extract and organize it on demand from any angle, which is 10x more efficient than search engines. Repeatedly asking questions from different perspectives and conducting cross validation using multiple LLMs can to some extent address the issue of AI hallucinations. As everyone says, LLM is the most patient, knowledgeable, and temperamental teacher we can encounter in our lifetime. The impact of this on social behavior is that, except for some people with high cognition and unique experience, we no longer need to chat with others about most issues. I don't want to see people who are visible or invisible. Except for friends. Stage 2: Skill Freedom The big change from LLM to Agent has abstracted SOP and unique experiences into skills that others can quickly replicate. · Mobile Internet restructured every industry, Crypto reissued every asset, and now AI is strengthening every profession (skill). Some expensive customized services in the past are being deconstructed by AI popularization, such as private doctors, which will bring about a round of service equality. When AI can provide basic medical advice with 80 points, the value of human doctors can shift towards the remaining 20% of complex/rare cases and emotional support. People with high agency can also play multiple skill roles simultaneously, dispatch different AI agents to complete the entire service/industry chain, and small companies with a few people can grow bigger. Before reaching the ultimate state of 'no one has to work', the gap between people (wealth gap) will be widened, not by 10% but by 10 times. Although various industries are already very focused on AI, AI agents are still in the early stages, and many professions have not yet felt the 10x productivity improvement brought by AI.