Own Your Solutions — Who Are Devs Without AI?
On the hidden cost of over-relying on AI, why some fear is rational, and what engineers at every level need to keep building for themselves.
Listen to under 2-min AI recap •TL;DR ↓
Own Your Solutions — Who Are Devs Without AI?
Engineers Must Own What They Ship
Writing code is becoming less central. But not irrelevant. Not yet. Not fully.
At the end of the day, we need efficient, concise, fast solutions — without going round and round delivering inefficiencies with AI-generated code. There's a cost to that: token spend, time, cognitive overhead. And sometimes the problem is just "change the colour of that button and this text."
Do you really need an AI agent for that?
The Hidden Cost of Over-Reliance
It's so easy to give up on thinking. To believe AI will solve everything on our behalf. But consider where this leads.
The real cost isn't tokens or API pricing. It's the slow, quiet erosion of the instinct to reason through your own code. Context-switching from your editor to a prompt and back. The cognitive offloading that slowly eats away at your ability to trace through your own codebase.
When the tool fails, changes pricing, or just gets it wrong — what then?
Some Fear Is Rational
I believe AI does not have to be feared. But with one caveat that matters.
Engineers who stop adapting, stop learning, stop being curious about the systems they build — they will fall behind. The ones who hand everything to AI and never look under the hood are building on a foundation they don't understand. When it cracks, they struggle to fix it.
The distinction isn't between those who use AI and those who don't. It's between those who keep sharpening their own thinking, and those who let AI replace the thinking entirely.
What we need to keep building is product knowledge, architecture knowledge, engineering judgement. The code itself may increasingly be generated — but the thinking behind it? That has to stay yours.
A Word for Juniors
It's never been easier to learn. You can ask questions without being judged. You can overcome shyness, the fear of looking stupid, the hesitation of asking for help. AI gives you that space.
But don't fall for easy vibe-prompting without trying to understand what it's doing. Use AI as a learning partner — ask it to explain, not just to generate. Look at the output and ask yourself if you could reproduce the logic without help.
You can learn by doing, iteratively, with AI as your most patient mentor. Just don't let it do all the thinking for you.
The tools will keep getting better. The question is whether we keep getting better with them.
TL;DR
- AI makes it easy to stop thinking — and that's exactly the trap; the real cost isn't tokens, it's the slow erosion of your ability to reason through your own code
- Engineers who hand everything to AI without understanding what it produces are building on a foundation they can't fix when it cracks
- The distinction isn't AI users vs. non-users — it's engineers who keep sharpening their judgement vs. those who outsource it entirely
- Product knowledge, architecture thinking, and engineering judgement are the skills worth protecting; code generation is just the easy part
- Juniors especially: use AI to learn and explain, not just to generate — it's the most patient mentor you'll ever have, but only if you push back on it
The tools will keep getting better; the question is whether you keep getting better with them.
Geddy
Senior Web Engineer / Lead
Engineering leadership • AI innovation • Product thinking. 20+ years of web engineering, from independent contractor to engineering leader. Passionate about developer experience and product engineering.