For the past two years, everyone has been asking the same question:
"Will AI replace software engineers?"
I think we're asking the wrong question. A more interesting question is
"What will software engineers spend their time doing when AI writes 80% of the code?"
I believe we're already seeing the answer.
Developers are gradually becoming AI code reviewers rather than code writers.
Recent studies show that developers using AI coding agents are merging more pull requests than before, suggesting these tools are changing how engineering teams work rather than simply replacing programmers.
Writing code is becoming the easy part
Ten years ago, our daily work looked something like this:
- Read requirements
- Write code
- Fix compiler errors
- Test manually
- Deploy
Today, an AI agent can generate hundreds of lines of code in seconds.
That's impressive.
But generating code was never the biggest bottleneck.
The real bottlenecks are still:
- Understanding vague requirements
- Finding hidden edge cases
- Reviewing architectural decisions
- Protecting security
- Keeping long-term maintainability
AI is getting better at the first step.
Humans are still responsible for everything that comes after.
The review loop is becoming more valuable than the coding loop
When I work with AI coding agents today, I spend far less time typing.
Instead, I spend more time asking questions like:
- Does this design actually scale?
- What happens if this API fails?
- Will another engineer understand this six months from now?
- Is the generated SQL actually safe?
- Did the AI silently change business logic?
Ironically, AI has made critical thinking more valuable than ever.
A great AI developer doesn't accept the first answer
One habit separates experienced developers from beginners.
Beginners often ask:
"Can you write this function?"
Experienced developers ask:
- - "What assumptions did you make?"
- - "What's the time complexity?"
- - "What are the tradeoffs?"
- - "Can we simplify this?"
- - "What could break in production?" AI responds to the quality of your questions.
Better questions produce better software.
The skill that's quietly becoming essential
I think one underrated engineering skill is AI review.
That includes:
- Verifying generated code instead of trusting it
- Spotting subtle logic bugs
- Evaluating architectural tradeoffs
- Improving prompts based on previous failures
- Knowing when not to use AI
These aren't "AI skills."
They're software engineering skills applied in a new workflow.
The future isn't AI vs developers
I don't think the next generation of great engineers will be the ones who type the fastest.
They'll be the ones who can:
- Design better systems
- Review AI-generated code efficiently
- Catch mistakes before production
- Build reliable guardrails around AI agents
- Turn AI into a force multiplier instead of a liability
AI can generate code.
But it still can't own the consequences of shipping bad software.That responsibility still belongs to us. And I don't think that's changing anytime soon.
Final thoughts
The rise of AI coding agents doesn't mean software engineering is disappearing.
It means the job is shifting.
Less time writing boilerplate.
More time reviewing decisions.
Less typing.
More thinking.
And honestly, I think that's a much more interesting future for our profession.
If you enjoyed this, you can follow my work on LinkedIn at linkedin
, explore my projects on GitHub
, or find me on Bluesky
Top comments (0)