AI and the Engineering Job
The honest version of what coding agents do to a career: which rungs of the ladder they remove, which skills they make more valuable, and what happens to the people who stop thinking.
6 articles · about 60 min in total
Start with AI Coding Agents and the Death of the Junior Developer PipelineThe honest version of this conversation is more specific than either the panic or the reassurance. Coding agents are genuinely good at a well-defined set of tasks, and that set overlaps heavily with what junior engineers were hired to do.
That is the actual problem, and it is structural rather than personal. The junior work was never only output. It was how someone became a senior engineer. Remove the rungs and the ladder still needs a top, but nothing is feeding it.
The second theme is what happens to the individual who leans on the tools without staying in the loop. The research on cognitive debt is early and worth reading carefully rather than quoting confidently, but the direction is intuitive: skills you stop exercising degrade, and the ones being outsourced first are the ones you need when the tool is wrong.
The way I think about it, the leverage moves to judgment. Deciding what to build, reviewing what came back, and knowing when the confident answer is wrong. Those get more valuable, not less, and that is the useful thing to optimize a career around.
Key takeaways
- Coding agents are strongest at exactly the well-scoped tasks juniors were historically given, which removes the rungs people used to climb into senior work.
- The skill that appreciates as generation gets cheaper is judgment: deciding what to build, reviewing output, and recognizing a confident wrong answer.
- Research on cognitive offloading suggests that skills which stop being exercised degrade, which matters most for the review skills an agent workflow depends on.
- Moving from software engineering to AI engineering is mostly a shift in evaluation discipline rather than in modeling knowledge.
Step 1: AI Coding Agents and the Death of the Junior Developer Pipeline
Employment for developers aged 22 to 25 fell about 20% from its late-2022 peak while developers over 35 kept gaining. The junior pipeline is breaking, and "just ship faster" doesn't fix the math of who becomes senior in 2032.
May 21, 2026 · 11 min read
Step 2: How I Use Claude Code and Coding Agents Without Becoming Useless
The engineers getting quietly replaced are the ones treating the agent like fancy autocomplete. A controlled study found AI made experienced developers 19% slower while they believed it sped them up 20%. The gap between those two numbers is where your job lives.
May 19, 2026 · 11 min read
Step 3: Knowledge Collapse and Cognitive Debt: Two MIT Papers on AI
Daron Acemoglu's new MIT paper proves there's an optimal AI accuracy, and it's not 100%. The knowledge-collapse equilibrium hinges on one parameter: how elastic human effort is. A separate MIT EEG study and three behavioral RCTs say it's elastic enough to worry about.
May 7, 2026 · 12 min read
Step 4: The Engineer-to-AI-Engineer Transition: What Actually Changes in Your Job
'AI engineer' isn't machine learning wearing a new hat. It's a different job that starts from a product and reaches for a model, not the other way around. Most job posts are confused about which one they're hiring for, and that confusion is where your leverage is.
May 17, 2026 · 11 min read
Step 5: Engineering Staffing in 2025: Hiring and Retaining in the Age of AI Augmentation
AI coding tools haven't eliminated engineering headcount, but they've changed what good looks like. Here's how to hire, retain, and level engineers in a world where the tools do more.
Jan 28, 2026 · 7 min read
Step 6: Frontend vs Backend vs Full Stack, Explained
Frontend is what you see, backend is what you don't, and full stack is doing both. Here is what each role actually builds, the tools they use, and honest career advice if you are just starting out.
Jun 8, 2026 · 8 min read
Frequently asked questions
- Will AI replace software engineers?
- It is replacing specific tasks rather than the role, and the tasks it replaces first are the well-scoped ones junior engineers were given. The near-term effect is on the entry path into the profession rather than on the total need for engineers.
- How do I use coding agents without losing my skills?
- Stay in the loop on the parts you would otherwise stop practicing: read the diff properly, keep writing the hard code yourself, and treat generated output as a draft from a fast colleague who is sometimes confidently wrong.
- What happens to junior developer hiring?
- The immediate pressure is real, because the work juniors were assigned is the work agents do most reliably. The unresolved problem is that senior engineers are produced by doing junior work, so reducing the intake creates a shortage several years out.
- How do I move from software engineering into AI engineering?
- Mostly by learning evaluation rather than modeling. The scarce skill is building systems that measure whether an LLM feature actually works, since the engineering around retrieval, tools and context looks a lot like ordinary distributed systems work.