The idea that software engineering is a career you can practice for 40 years may be ending. The reason is not that AI will replace all programmers tomorrow, but that the economics of the profession are shifting in a way that mirrors professional athletics: a short, high-earning window followed by forced retirement.
The core argument
Sean Goedecke, a software engineer and writer, argues that the rise of AI code generators creates a fundamental tradeoff. Using AI to write code means you learn less from each task. Over time, your technical skills may atrophy. But that doesn't matter for your career prospects, because refusing to use AI will simply get you outcompeted by engineers who are willing to make that trade.
Goedecke draws a direct analogy to construction workers. Lifting heavy objects damages your back and joints over the long term, but the job requires it. You don't get to say "I'm a good construction worker, so I won't lift heavy things." You lift them, or you find another line of work.
The athlete analogy
The most striking comparison is to professional athletes. A pro athlete's career typically lasts about 15 years. You make a lot of money until your mid-thirties, then your body gives out. The tragic figure is the athlete who doesn't prepare for that day.
Goedecke suggests software engineers may be in the first generation to face a similar trajectory. If AI tools continue to improve, the premium on human coding skill may decline sharply after a certain age or experience level. The engineers who plan for this — by saving aggressively, building alternative skills, or preparing for a second career — will be better off than those who assume the old model will persist.
What this means in practice
This is not a prediction that all software jobs vanish. It's a prediction that the career structure changes. The old model — where you could turn a coding hobby into a lifelong, increasingly lucrative career — was never an immutable fact. It was a fortunate coincidence of the industry's growth phase.
If AI does make engineers less effective over time (and Goedecke notes that the evidence is not yet clear), the profession may still be obliged to use it, because that's what employers will pay for. The alternative — refusing to use AI and writing code by hand — may become economically unviable, much like a carpenter refusing to use power tools.
The Luddite comparison
Goedecke also references the historical Luddite movement, where 19th-century weavers destroyed machines that automated their work. Some modern anti-AI activists have reclaimed the term, arguing that resistance to automation is justified. But Goedecke is pessimistic about tech labor unions or collective action slowing the process, citing the high pay, remote work possibilities, and global competition that make organizing difficult.
Bottom line
The takeaway is practical: if you're a software engineer, plan for a career that may not last 40 years. Save money, develop skills outside of coding, and be realistic about the trajectory of the tools you use. The profession is changing, and the engineers who acknowledge that change will be better positioned than those who pretend it isn't happening.