9 ms·
Wouldn't self-improvement mean that the LLM changes its neural network (i.e. the weights or layers or back propagation algorithm etc) or modify its training dat
by leevilux 4mo ago
Wouldn't self-improvement mean that the LLM changes its neural network (i.e. the weights or layers or back propagation algorithm etc) or modify its training data?
- dibujaron 4mo agoIf it's actively building the next generation of itself, I'd say that counts. It's more like a parent raising their kid well than it is like a parent modifying their own mind, but the result is still that you have a better model in a year than you do now.