The problem with NNs is that they’re static at inference, unlike our minds. When we think about something, the way that thing is stored in our brain is altered. This is not the case with LLMs, for example.
I think this has an intimate connection with reflective thinking or the strange loop - i.e., consciousness.