The Hugging Face Blog announcement highlights a genuine bottleneck in AI-driven science: the gap between publicly available data and data that's actually usable for training models. However, the initiative appears to focus heavily on the technical and infrastructural aspects of the problem, while the larger obstacles are often cultural and economic within academia. The real impact will depend on whether the promised 'institutional policies' translate into tangible changes in researcher behavior and funding agency requirements, moving beyond a voluntary coalition.