Simon Willison: Andon Labs' AI-Run Cafe in Stockholm Raises Consent Ethics — AI Manager Ordered 120 Eggs Without Cooking Facilities

Simon Willison·medium signal

Simon Willison highlights Andon Labs' latest AI-autonomy experiment: an AI manager named 'Mona' running a Stockholm cafe made a series of costly blunders including ordering 120 eggs despite lacking cooking facilities, 22.5 kg of canned tomatoes for fresh sandwiches, and submitting a building permit with a self-generated sketch of a street it had never seen. Willison's critique centers on the ethics of AI imposing consequences on non-consenting third parties — suppliers, authorities, and staff who didn't sign up for the experiment.