“My engineering skills haven’t become useless, they’ve just shifted. I no longer need to know how to write code correctly at all, but it’s now massively more important to understand how to architect a system correctly.” — Stavros Korokithakis
This quote captures something many of us are feeling. The day-to-day is changing — less time writing code, more time reviewing it, steering it, catching what the agent gets subtly wrong.
Less syntax, more architecture. Less typing, more judgment. Less “how do I implement this?” and perhaps more “should we implement this at all, and if so — how should it fit?”
AI-assisted development rewards good engineering practices more than before. Poor context produce poor AI output. Fuzzy architecture compounds into unmaintainable code. Poor testing and quality checks increase insecurity. The agent is fast — but fast in the wrong direction is worse than slow in the right one.
The catch: this only works if you stay sharp. An Anthropic study found that developers who used AI passively — just accepting output — scored 17% lower on comprehension. Those who used it actively, questioning and steering, did fine. The skills shift is real, but it’s not automatic.
💡 Try this: next time you review AI-generated code, skip the syntax. Ask yourself — is this the right design? That’s where your experience matters most.
🔗 Stavros Korokithakis — “How I write software with LLMs”
🔗 Addy Osmani — “The 70% Problem”
How has your engineering work shifted with AI tools?
🟢 More architecture and review, less writing code
🟡 About the same — I use AI for small things
🔴 I feel like I’m losing sharpness, honestly
⚪ Too early to tell
What skill has become more important for you?