2026-10-29 21:02 UTC
If it makes you feel in better, I think models are “topping out.”
They are getting better and better and writing functions, and even planning some structure.
But, fundamentally, autoregressive LLMs are not good at architecting large projects sustainably. It is just a terrible architecture for that, and they aren’t getting around that; it’s like getting a weather prediction ML model to write novels. It’ll take an entirely new class of systems.
Replies (1)
-
@red_tomato@lemmy.world 2026-10-29 23:42
I doubt models are “topping out “. It feels like no prediction about the supposed “fundamental limits” of AI models have held. Two years ago it seemed like LLMs were fundamentally incapable of doing maths. It’s just a next word predictor! Of course it can’t do maths! Now it’s solving high profile maths problems left and right faster than any mathematician alive. Still a next word predictor.