|
|
|
|
|
by sometimelurker
32 days ago
|
|
I think just "semblance" would be easier. anyways, I'm pretty sure there's nothing stopping the development of an intelligent system. I do also think that pure "dumb" scaling of today's LLMs can bring that. They can develop "understanding" of simple things (multiplication), and just by making them larger they can "understand" other things (but only if the training process incentive this) |
|