Ask HN: Why do these LLM models seem to improve like clockwork over time?
By iknownthing · 2026-11-22 · 2 points · 2 comments
It seems like these AI labs are able to guarantee more advanced models over time. How are they able to do this? Are they continually finding better architectures and learning algorithms or maybe even better data? Or are they simply making the models bigger?
Open the full discussion on BetterNews