> The claim that LLMs are nothing more than fancy expensive markov chains is a dangerous oversimplification or, in other words, wrong.<p>But they are fancy expensive Markov chains. The learning here is that fancy expensive Markov chains can do a lot, not that the statement is wrong.
I agree with the general point that there is, at present, far too much uncertainty to understand the long-term effects of LLMs. It will be somewhere on the scale between "semi-useful toy," "ends white-collar work as we know it," and "accidentally destroys mankind." Beyond that, there isn't much we can say.
If physical reality is quantized, which many believe to be true, then the state of all all reality can be described symbolically. Therefore, a large enough language model could conceivably model physical reality. At some point, we could construct a LLM powerful enough to manipulate the lowest level quanta of reality given a description of the manipulation in another language. The tricky part is resolving ambiguity, as the only truly unambiguous description is the direct description of the changes at the sub-atomic level.<p>What is the language of the human mind? Of consciousness itself? If a LLM can learn that, perhaps we will build a technology that takes an intention and is immediately able to manifest the corresponding changes to reality to satisfy the intention or desire? When we get there, perhaps it will be time to look inward and do some of that spiritual inner work we keep putting off?<p>What’s exciting to me about LLMs is they seem to be one step closer to this vision, with all the peril and all the possibility that comes with it.