Advice to a beginning software engineer
Short summary
A senior engineer argues that junior developers should ignore idealistic advice from the ZIRP era and instead focus on being useful, friendly, and conscientious while learning to use AI as a tool without delegating their own judgment to it. The core message is pragmatic survival: build value through competence and likability rather than fighting political battles you can't win.
- Much current advice from senior engineers is outdated ZIRP-era thinking that assumes abundant investment money and engineer leverage; junior engineers lack the political capital to act on it without getting punished.
- You should use AI tools because your company expects it and you have no leverage to refuse, but never delegate your actual judgment to them—ask questions, disagree when warranted, and understand what you're passing along to others.
- Being conscientious and helpful covers more ground than trying to play office politics or take stands; focus on deep understanding of the systems you work with to build genuine technical value.
Full summary
The core argument here is that most advice floating around tech right now is outdated because it was shaped during the ZIRP era—roughly 2010s to early 2020s—when money was cheap, companies competed aggressively for engineer talent, and you could actually push back on work assignments or company politics without risking your job. That world is gone. Goedecke warns junior engineers away from senior engineers telling them to unionize, fight "unethical" technologies, or insist on doing work their own way. That advice might have worked when you had real bargaining power; now it's a fast way to get fired if you don't have leverage. The real play for beginners is different: be useful, stay friendly, keep your head down, and don't start political fights you can't win.
Instead of playing office politics or making stands, Goedecke recommends building value through conscientiousness—asking questions, actually understanding the systems you work with instead of assuming someone else has it covered, and becoming the person who knows details nobody else bothers learning. This generates real technical value without requiring you to take risks. He also pushes back on the idea that you should avoid AI entirely. That's ZIRP-era thinking too. Your company will expect you to use it the way construction expects workers to use power tools. But here's the critical part: don't become a "meat proxy" who just forwards AI suggestions without thinking. Ask questions, disagree when you disagree, and use AI to inform your own judgment rather than replace it. Panic about being obsolete is what drives people to abdicate responsibility to models.
The final point is straightforward—don't lose hope, but also don't panic. Technology this disruptive always has unpredictable knock-on effects, both terrible and amazing. People have predicted the end of software engineering before (outsourcing, high-level languages, low-code tools) and been wrong. What won't change is that being smart, friendly, and conscientious will matter. The job might transform, but deferring all your thinking to an AI in the name of despair is a self-fulfilling prophecy that actually makes you less valuable, not more.