"o1 is still flawed, still limited, and it still seems more impressive on first use than it does after you spend more time with it," he wrote. "[B]ut also, it is the beginning of a new paradigm: AI that can do general-purpose complex reasoning."
You are viewing a single comment's thread from: