Such a crazy change from the times of Luminous, when they published a three pager with a claim that the model is similar good as „gpt 3“ (which??) with some graphs without y axis.

Bravo team!