I think this kind of thing is a pretty strong argument for the entire open source model ecosystem, not just open weights but open data and the whole gamut.