The Illusion of Competence

Summary

Technologists Needs More Than AI Proficiency

Over the past five years, a combination of AI and Covid-era learning gaps, has resulted in a dramatic shift in how young engineers learn and build. AI copilots now write functions, explain code, generate APIs, fix bugs, produce documentation, and design entire architectures in minutes.

It’s intoxicating.
A new developer can produce working software faster than any generation in history.

But, we are creating a generation of developers who can build things that _may_ work, without understanding why they work — or when they will fail.

Or in simpler words: They have been imparted ‘knowledge’ without the foundation to gain ‘experience’

Managers quietly share the same concerns:

Engineers who can code, but can’t debug.
Builders who can assemble components, but not explain tradeoffs.
Programmers who trust AI answers even when the answers are wrong.

This is not their fault. They never had to learn slowly. Organizations have an obligation to help them.

The Compass: The Original “AI Problem”

Seven centuries ago, sailors faced a similar trap.

When the magnetic compass spread across Europe in the 1200s, it felt like magic.
No stars required. No experience needed. Just follow the needle, and you’ll arrive safely.

Except — the needle didn’t point to true north.
It pointed to magnetic north, which shifts by geography and time.

Early navigators didn’t know this.
They trusted the tool, not the science.

So ships left port confidently…
and drifted degree-by-degree into storms, reefs, hunger, and disappearance.

The compass wasn’t the enemy.
Blind reliance was.

It took centuries — and William Gilbert’s De Magnete in 1600 — for humans to realize the Earth itself was a magnet, and that variation had to be calculated, corrected, and taught.

Only when sailors understood the tool
did the tool become safe.

Knowledge was knowing how to use the compass.
Experience was knowing when the compass could get them killed.

AI Is Today’s Compass

AI points us forward — fast, confidently, convincingly.
But a correct-looking answer is not a correct answer.
A working prototype is not a durable system.

If we allow students and early-career developers to trust every output, bypass fundamentals, avoid learning what breaks and why, then we are sending ships to sea without navigators.

We Need More Than Tool Training

It’s not enough to teach young engineers how to use AI.
We must teach them how to challenge it, validate it, and override it.

That means a structured way to build using engineering intuition, independent reasoning, skepticism and healthy dose of craft knowledge

A Practical Solution: The “Dual-Track Engineer” Program

A repeatable model for universities and companies.

Track A: Assisted

Students build systems with AI tools:

Plan → Code → Test → Deploy with AI support
Optimize speed and outcome
Learn prompt engineering and tool capabilities

Track B: Unassisted

Then they rebuild critical parts without AI:

Write core logic by hand and Solve bugs without LLM help
Challenge themselves to review LLM code and suggestions to improve outcomes
Design architectures independently
Explain performance and reliability choices

The Bridge: Tool Reconciliation Report

For every module, they answer:

What did the AI produce?
What was wrong or suboptimal?
What did you change, and why?
What did you learn that the tool could not explain?
Could you do it without the tool? If not, what’s missing in your knowledge?

This framework develops judgment, craftsmanship, architectural thinking, and the ability to spot failure before it happens.

Organizations need to drive this change

Sailors became true navigators only when they understood the compass deeply enough to doubt it. Tomorrow’s engineers need that same mindset.

Learn the fundamentals.
Question the output.
Understand the why, not just the what.

AI is a powerful compass.
But the world still needs navigators.

Knowledge is knowing what works.
Experience is knowing what doesn’t.
And technology moves forward only when we have both.

Read about the story of Apollo-11 and how Neil Armstrong took control when the ‘automated’ systems failed