More of a scaling issue: Humans do continuous* online learning, while LLMs get retrained once in a while.

* I'm no expert, 'continuous' might be oversimplified.