$Tesla(TSLA.US)From a technical perspective, why is FSD difficult to implement? The current stage of AI is based on statistical principles, which means enumerating as many scenarios as possible and then fitting countless functions to explain the data. Neural networks are essentially the organizational structure of these countless functions. This type of AI lacks true understanding. Its nature dictates that with enough data, it can easily surpass humans in highly specialized fields. However, conversely, for complex scenarios without sufficient data, it cannot fully solve the problem. This is the limitation of current AI models—no real intelligence, not enough time, and insufficient data mean FSD can never surpass humans.