Depends on the goals. If they were fine-tuning leading foundation models, then I could see this being an entirely sensible undertaking. But since their goal seems to be to make foundation models, I don't think that they will end up being the leading models with so many other conflicting requirements.