Several of today’s most interesting AI stories are not about raw capability; they are about whether people want AI around them at all. Librarians are teaching people to avoid AI, Lorde is rejecting AI glasses as a vibe, Anthropic’s ad is creeping people out, and OpenAI’s hardware push is testing whether a dedicated AI device can feel useful rather than invasive. For product builders, the message is blunt: consent, taste and reversibility are becoming product features.