What if it could learn and update its own weights as it's being used?

The model becomes uniquely yours, no KV stashing, no MD files, no trickery.... just live learning.

Unload the model, move it to another machine, and it still knows what you told it. Could this be the start of recursive self improvement for local AI?