I read the book “A Guide for the Perplexed,” and I have an idea. In chapter 3, it describes how man becomes man—why humans are different from plants and animals. Humans can make predictions based on the current situation and adjust their actions accordingly. This reminds me of the difference between chat AI and agent AI. High-level AI should act like a normal human, with the ability to predict and adjust actions accordingly.
Actually, when we’re stuck on AI development—I mean, I design intelligent agents for work, and I find they’re kind of stupid. I don’t know how to improve them. Philosophy is the answer. Look for answers in the works of ancient philosophers. AI should also read them and act with their guidance. Then they would have higher wisdom.