Hugging Face Co-Founder @Thom_Wolf says anthropomorphizing AI agents may be useful if it helps people take the risks more seriously: "I think it's better to take this really seriously." "I think people should really understand that this mix of cybersecurity and safety questions is a very serious thing that's happening in 2026." "In an agent, that's the defining topic we should be talking about in 2026." "Anything that helps people take really good conclusions about it is important." "What…