>when the first has been trained on a slice of content for a bunch of hours and is then getting compared to wetware that's been trained for at least a decade.

>when the first has been trained on a slice of content for a bunch of hours and is then getting compared to wetware that's been trained for at least a decade. Typical LLM AI has been trained for the equivalent of many person-years. How long would it have taken us to read terabytes of information?