The era of scaling is over, research is back

A new essay expands on Ilya Sutskever's claim that the era of simply scaling AI models has ended, marking a return to the era of fundamental research.