Comment by s0ulphire

A follow up question: as a human doesn't start with "knowing" something either and first creates definitions for objects or words, which it then uses to build increasingly abstract concepts that we eventually classify as "knowledge" on the thing, is there anything that would stop LLMs from being able to do the same thing? I fully agree the capability is not there yet, but I can't say what would stop an appropriately designed model from being able to do so myself.