Junior devs — typing to delegating¶
March 22, 2026 · LinkedIn
20 reactions · 6 comments · 1,447 views
Lately I've had many conversations with developers who are worried.
Two groups, same worry — just different flavors.
The newly hired consultants who want to learn the trade. The students who haven't finished their degrees yet. Both asking some version of: what do I actually need to learn now?
I told them I don't have all the answers.
But I said this: every previous technology shift changed the tools. This one is changing the nature of the work itself. Not from one language to another. From typing to analyzing and delegating.
That is a harder transition for someone who is still learning to type.
The honest answer I gave: the uncertainty is real. But so is the opportunity. If agents can execute across domains, the practitioner who knows how to direct them gets to experiment much wider than before. The T gets broader, and deeper, faster than any previous generation could manage.
That's not a guarantee. But it's the frame that feels most honest to me.
What are you telling junior developers when they ask?
Discussion¶
I usually tell junior developers that the shift to AI hasn’t reduced the importance of fundamentals—it’s amplified them. Writing code was never the real differentiator understanding problems, designing systems, and making good decisions was. Now that execution is faster, the real edge comes from how well you can frame problems, validate outputs, and think through trade-offs. The engineers who will thrive aren’t the ones who type the fastest, but the ones who can guide, question, and orchestrate systems effectively.: I usually tell junior developers that the shift to AI hasn’t reduced the importance of fundamentals—it’s amplified them. Writing code was never the real differentiator understanding problems, designing systems, and making good decisions was. Now that execution is faster, the real edge comes from how well you can frame problems, validate outputs, and think through trade-offs. The engineers who w...
Thor Henning Hetland: Thor Henning Hetland That’s very well put the illusion of fluency is probably the biggest trap right now. It’s easier than ever to get something working, but much harder to know why it works or where it will break. I think the discipline now is to deliberately slow down at key points trace the system, question assumptions, and build mental models alongside using the tools. Otherwise, as you sai...
Totto ↩: Himanshu Verma That's the trap in a single sentence. And it scales to teams too, not just individuals. The velocity is real but the debt is quiet. What I've started doing: explicit checkpoints after sessions — not to review the output, but to ask what I actually understand now that I didn't before. Treating comprehension as a deliverable, not a side effect.