Fair point. But you can be correct and petty at the same time. Remember that he blamed the people using Galactica casually as the reason it got paused. Then wonders and asks people why ChatGPT hasn't faced the same backlash given that "it spouts sh-t*.

Although one could argue that usable LLMs in production are quite revolutionary. NVIDIA'S GauGan or GAN based txt to image models, the base diffusion models have been there for a year or two but hasn't received the same publicity and profits as Stable Diffusion or Midjorney. It's basically the same line of framework.

It's narrow-minded thinking to brush the architecture upgrades and the engineering work that made it possible -- which has always been his statements. But that is a fair point considering he is mainly a researcher not an engineer.