Y
Hacker News
new
|
ask
|
show
|
jobs
by
chikere232
523 days ago
> Technically, the models can already learn on the fly. Just that the knowledge it can learn is limited to the context length.
Isn't that just improving the prompt to the non-learning model?