The closer you get to AI, the more replaceable you become

"The real job of a senior engineer is to make more senior engineers." Scott Hanselman, a VP at Microsoft, made this observation in response to a pattern he was watching: many tech companies have slashed junior engineering positions, letting AI handle the entry-level work. Headcount drops, efficiency metrics look cleaner. The logic is straightforward. The problem comes later. Junior engineers aren't just doing grunt work. They're also following a path that, years from now, leads to becoming senior engineers. Cut that path, and who handles the complex judgment calls, the aging legacy systems, the next generation of engineers coming up behind them? AI is a better tool. Using it doesn't transfer away the responsibility to develop people. A lot of companies seem to be treating those two things as interchangeable. A court case in Hangzhou, China, shows a different version of the same problem. An engineer known as Zhou You joined a tech company in 2013 and was eventually assigned to AI large model quality inspection: reviewing AI-generated content, catching errors, keeping the output reliable. Early this year, he received a termination notice. The stated reason: economic headwinds, organizational restructuring. He filed for arbitration, claiming wrongful dismissal. In court, the company admitted the real reason: the AI had improved to the point where it fully handled his role. The judge asked for documentation showing how the substitution was approved and planned. The company had nothing. Both instances ruled against them. Zhou You was awarded over 260,000 yuan in compensation. His lawyer put it plainly in court: "This job only existed because AI needed it. He spent his days correcting the AI. In the end, he corrected himself out of a job." Two situations. Two different levels. The Hanselman case is about companies: cutting junior positions saves visible costs, while quietly eliminating the path for growing senior engineers. Zhou You's case is about individuals: positioning yourself as an AI specialist, working as close to the technology as possible, offered no particular protection. Both logics hold on their own terms. Neither resolves the other. As of now, no one has laid out a clear path for handling both at once.