>>519199503
The current batch of LLM's DO NOT learn like humans, in that they do not retain factual datapoints, but instead retain only context weighted probabilities. LLM's have got something like a slice of our associative memory system and language processing, but they're not close to a full brain on their own.