Knowing human history and how the present came to be, we can answer this confidently by stating that what is convincing people that chatbots understand and can reason, and that AI is doing an increasingly sophisticated and successful representation of human thought, is a kind of mind control manipulation to begin with. Here we get back to the limits of language. As a representation of language, a good frame of reference in this discussion is represented in the bottom line from a famous poem: "Poems are made by fools like me, but only God can make a tree." ["Trees" by Joyce Kilmer] ChatGPT or Grok or any of the other AI systems under development can talk about trees from many different perspectives and vantage points, even write poetry about trees, catalog the species perhaps, and regurgitate much seemingly hard science which will inevitably be surface knowledge because there is no handbook in human existence for how to make a tree in actuality. The rest is all talk and no action. So we are trying in this way to show you that AI is all about surface descriptions, surface representations, partial knowledge, partial understanding, and limited utility in the kinds of knowledge it can assist with and the subsequent benefits. It is because you are living superficial lives to such a high degree that a fancy new way to create a simulation of living is impressive to you and compelling, but we see it as a semblance of the real thing, always, no matter how complex and seemingly sophisticated it might be as a description or even an artistic representation.
So the question then, simply comes back to how much value is placed on simulated existing compared to something real. To have a real-world application and real-world significance so you learn something from AI that can be translated into physical action in some fashion to gain an advantage over nature or a way to manipulate physical reality that gives you a material benefit is creating value. The question is to what degree that can be engineered to happen reliably, and at what cost if it requires a gigantic investment of capital to create a semblance of reality and in the creation of so many individual iterations, sorting through, ranking them, and having a way to know you have the best of the best may not even be feasible. So the problem will be one of knowing when you have truly made an advance and when you have been misdirected to have a kind of simulation of progress but not a true advance. There is a truism at work here that will be appreciated more and more with this kind of endeavor being pursued so avidly by so many, that "large gains will have a correspondingly large expense and large risk associated." It is much like the fundamental problem we started with in this discussion, that one could have asked at the outset whether, given the current state of knowledge and what is wanted to happen, is it even feasible to do so without exceeding available resources in terms of the actual cost involved. That cost is real and mounting; it could easily use up the energy of the planet and increasing numbers of the time and talents of human beings, and the unanswered question will always be whether it is worth it in the end.
Please login or Register to submit your answer