What is the status on continual learning for LLMs?

5 points by fred123123 ↗ HN
I belive if this is solved, agi is very close. So id like to know if its solvable

3 comments

[ 7.8 ms ] story [ 30.7 ms ] thread
The issue remains the context window. I haven't seem any example of a "memories" system that the agent can update over time without it devolving into slop. The magic of the human brain is that we seem to have decently sophisticated heuristics parsing/saving memories as well as for letting memories that aren't accessed decay.
It depends on what you mean by continual learning. Do you just mean some form of memorization, or something more/different?