Deterministic deductive logic can be easily implemented using computationally comprehensible algorithms. However, uncertain inductive logic cannot be directly implemented by computers because computers lack perception. So, who can act as the perception engine for machine intelligence? The answer is: humans. AI cannot progress without humans. Only humans can help AI perceive reality. For example, the inductive organization of large-scale model data is done by humans and then handed over to machines for training. For example, the consensus value of Bitcoin comes from people's perception through buying and selling $BTC, forming a sea of price fluctuations that allows the consensus on Bitcoin's value to emerge.