This is the inherent conflict in the company. Once it turned out you needed 7B parameters or more to get LLMs worth interacting with, it went from a research p

This is the inherent conflict in the company. Once it turned out you needed 7B parameters or more to get LLMs worth interacting with, it went from a research project to a winner-take-all compute grab. OpenAI, with an apparent technical lead and financial access, was well-positioned to win it. It is/was naive of the board to think this could be won on a donation basis.