AI isn't just eating jobs. It's eating the career path.
Diesen Artikel auf Deutsch lesen
There’s a comfortable lie in the tech world.
That AI makes us more productive. That it takes work off our hands. That it frees us up for the “important things.”
That’s true.
And at the same time, something much more uncomfortable is also true:
AI is eating the path we use to become competent in the first place.
Developers always thought they were outside the problem
For decades, the roles were clear.
Developers build systems. Others lose jobs.
Automation was something that happened to other people.
And the reactions were always the same:
- “That can’t replace real work”
- “Quality will suffer”
- “It doesn’t understand reality”
Today, developers are saying the exact same things.
Only this time, it’s about their own job.
This isn’t a technological shift. It’s a perspective shift.
AI is not creative. It is ruthlessly efficient.
The romantic idea of software development is quietly dying.
Code as craft. Elegance. Artistry.
AI doesn’t care about any of that.
AI cares about:
- outcomes
- speed
- scale
And suddenly it doesn’t matter how you build something, only that it gets built.
That changes everything.
From doing to directing. From writing to deciding. From junior doing to senior thinking.
And this is where the real problem starts
If one senior with AI can do the work of five juniors, why would a company still hire juniors?
There is no moral answer to that. Only one that works in a spreadsheet.
And it’s brutally simple:
They don’t hire them.
This isn’t a future scenario. It’s already happening.
The mistake isn’t in the present. It’s in the future.
Senior developers don’t appear out of nowhere.
They are built through:
- mistakes
- bad decisions
- context
- years of “not knowing what you’re doing”
That’s exactly the part AI is removing.
We are optimizing so hard for efficiency that we are removing the inefficient part that actually creates expertise.
It’s like getting rid of apprentices because masters work faster.
Great in the short term. Fatal in the long term.
We are building a system with no pipeline
The result is paradoxical:
- companies want senior developers
- while eliminating the path to become one
A system that only needs experts but produces none is not efficient.
It is unstable.
And then comes the uncomfortable question
Who is still responsible for training?
The answer used to be clear: companies.
Today, it isn’t.
Because companies have an incentive not to do it anymore.
And that leads to the real question:
Who pays for the missing education?
We are not going to like the answer
There are only three options:
The individual pays: more self-optimization, more pressure, more inequality
The state pays: companies save money, society carries the cost
No one pays: and we notice only when the experts are gone
This is not a technical problem. It is a distribution problem.
The most beautiful irony of all
For years, developers complained about the business side:
- requirements are unclear
- specifications are poor
- “the user doesn’t know what they want”
Now those same developers sit in front of an LLM and write prompts.
And suddenly:
- unclear requirements
- contradictory inputs
- outputs that are “technically correct, but wrong”
The roles have flipped.
The developer has become the business side. The LLM is the engineering team.
And the complaints are identical.
The real problem is not AI
AI works.
The problem is what it quietly removes:
- learning curves
- friction
- mistakes
- time
All the things we hate. All the things we need.
We are optimizing ourselves out of our own future
This is the pattern emerging:
- short term efficiency wins
- long term capability loses
Or put differently:
We are building a system that works perfectly as long as it lives off the past.
The uncomfortable truth
We always believed technology replaces work.
This time, it replaces something else:
the path to capability.
The real question
Not:
“Is AI good enough?”
But:
What happens to a society where no one has to take the long path anymore and therefore no one truly can?
Conclusion
AI may not be the end of developers.
But it might be the end of what makes a developer.
PS: The irony described here is hitting developers hard right now, and it feels new because until now it always hit others first. But this is not new. Clerks, designers, translators, journalists. In all of those fields, the path from beginner to expert has already been quietly eroded. The only reason developers are noticing now is simple: it’s our turn.
Further reading: