https://towardsdatascience.com/gpt-4-is-coming-soon-heres-what-we-know-about-it-64db058cfd45


Altman said they weren’t focusing on making models extremely larger anymore, but on getting the best out of smaller models. OpenAI researchers were early advocates of the scaling hypothesis but may have now realized other unexplored paths can lead to improved models.


스케일링만으로는 AGI 불가능하다고 보고 개선된 훈련회수나 알고리즘쪽에 집중하려는거같다