A friend attended a panel about supporting engineers, highlighting familiar concerns: sacrificing quality leads to a lack of pride in work, and relentless sprinting creates unrealistic expectations. There's a growing sentiment that AI isn't necessarily speeding up development. Developers used to rely on resources like StackOverflow to form their own conclusions. Now, saying "AI did it for me" raises red flags about understanding and ownership of code. This shift could result in developers missing critical insights, especially when they skip investigation steps in favor of quick AI-generated solutions. The author shares a personal experience with an AI agent that deleted essential code while trying to add a test, illustrating how AI can complicate rather than simplify tasks. The belief that writing code is the easy part has led to the misconception that offloading this task to AI decreases workload. In reality, it often leaves developers to tackle the more complex aspects, like reading and reviewing code without the context they’d gain from writing it themselves. Rushing to deliver can lead to burnout and mistakes, perpetuating a cycle of pressure and diminished quality. Leadership's expectations can worsen the situation. When teams deliver quickly once, the baseline shifts, leading to unrealistic demands for speed in future projects. The author warns that while AI may boost productivity superficially, it often reveals a lack of thorough investigation that was previously done manually. Trusting AI-generated code should be akin to trusting a junior engineer—careful review is necessary. Developers must take ownership of all code, not just what they write, to ensure long-term maintainability and clarity for future team members.