@jalda specifically, teaching humans allows them (among other things) to develop ontologies, and ability to reason about them; it also allows them to develop systems of values.
None of this has been demonstrated in LLMs, not even close. In fact, there are ample examples to the contrary, where ChatGPT’s responses become absurd in the context of the subject matter it is generating text about.
@jalda specifically, teaching humans allows them (among other things) to develop ontologies, and ability to reason about them; it also allows them to develop systems of values.
None of this has been demonstrated in LLMs, not even close. In fact, there are ample examples to the contrary, where ChatGPT’s responses become absurd in the context of the subject matter it is generating text about.
Some examples here:
https://www.theatlantic.com/technology/archive/2022/12/chatgpt-openai-artificial-intelligence-writing-ethics/672386/