▲ 1 pointsCoercion and Deception in AI-to-AI Management: An Agentic Benchmarkarxiv.orgby sbulaev·27d ago·0 comments·view on hn ↗