Yeah. I was also in Paris for the AI Action Summit, where we launched the “ROOST,” the Robust Open Online Safety Tools. Part of the ROOST idea is that the defense against attacks that are clearly unlawful should be open source and decentralized as well. For example, child sexual abuse material is illegal to host. This is not a matter of lawful or awful speech; it’s basically illegal content. Currently, we rely on Big Tech to tell us whether material is the same as known child sexual abuse material or not. Smaller operators have to pay very large licensing fees. If there are errors in the judgment, there’s really no way to appeal outside of existing Big Tech procedures.