My model of what is going on with LLMs
People seem to be confused about what can be concluded about scaling large language models. At this point, it is not plausible to precisely predict how far we are from unlocking all of the remaining functions that large language models are capable of. We don't even know all of the component functions of human intelligence - every time we solve something previously out of reach, it turns out that human-level generality is even further out of reach. The invention of human-level artificial general intelligence will disrupt history, but that point is still pretty far off.