Comment on The model is the productparentComments−bjornsing1yThat doesn’t make sense though if scaling is actually stalling. The reason so much compute goes into training now is scaling, which keeps base model lifetime short.
Comments
That doesn’t make sense though if scaling is actually stalling. The reason so much compute goes into training now is scaling, which keeps base model lifetime short.