We’re living through a strange era of work: leaders are told AI enables “10×” or “100×” productivity, while engineers are quietly fixing AI-introduced bugs, validating hallucinated outputs, and absorbing the mental load of tools that are powerful, but deeply imperfect. This talk is about surviving that reality. We’ll examine what actually happens when AI enters real developer workflows: where trust is partial, verification is mandatory, and speed often comes at the cost of focus and creativity. We’ll talk candidly about productivity theater, unrealistic expectations, and the quiet burnout that happens when “just use AI” becomes a substitute for good leadership and sustainable engineering practices. Drawing from real conversations with engineers and teams, this session offers practical ways to: • Push back (constructively) against impossible productivity narratives • Use AI without surrendering judgment, craft, or accountability • Protect mental energy in environments obsessed with output metrics • Re-anchor DevOps values around humans, not hype cycles This isn’t a talk about how to get more done. It’s a talk about how to keep doing good work, and maintain health and stability, in a world that thinks software engineers should be machines.