Our position on open-weights models

This looks like a summary. Want the full article?