Что может сделать ваше тело , чтобы ваш стул никогда не мог повторить?
Конечно, при конкретной задаче « держать вещи и людей комфортно » ваш стул намного лучше, чем ваше тело. Как насчет буквально любой другой задачи - например, производить эритроциты?
Today’s AI is task-specific. Is the number of “tasks” finite or infinite? Okay, let’s assume there’s an AI algorithm for every possible human “task” and call that massive chunk of algorithms “AI”, going forward. Here are some questions about the human brain that have been discussed for thousands of years, but are still meaningless for the glorified logistic regression that is today’s AI:
Asking these questions with respect to today’s AI is like asking “What life goals does this couch have?” or “Does this rock dream?”. They are meaningless.
Major functions of the human brain - functions which make a human human - like consciousness, creativity, morality, empathy, self-awareness, emotion, ambition, hope, etc. simply have no parallels to today’s AI, unless you radically change the definition of any of those terms.
Of course, there’s certainly a possibility that there will be a paradigm shift in the field of AI in the future, where every single electrochemical reaction in the human brain can be reliably mapped and empirically simulated in an algorithm which people usually call AGI - Artificial General Intelligence. This will first of all, require a fundamental understanding of the human brain, of which we now knowAnd then manifesting every little detail into an algorithm - one which will probably look nothing like today’s AI algorithms. Until that is achieved, comparing the brain and “AI” is like, to quote Andrew Ng, “talking about over-population on Mars”.
It’s simply a non-starter before we even reach Mars (AGI).