We Share Everything. Almost.
Engineers have the strongest knowledge-sharing culture of any profession.
We contribute to open source. We write technical blogs. We speak at conferences. We review pull requests line by line so a junior doesn't ship the same mistake we made three years ago. We write READMEs, CONTRIBUTING.md files, and detailed issue responses — all so the next person doesn't have to suffer what we suffered.
This is the culture we should be proud of.
But there's one thing we're not sharing.
How to think with AI — not just how to use it.
The Structural Reversal No One Talks About
Every previous technology wave — PCs, the internet, mobile, cloud — favored the young. Younger generations adopted faster, built faster, disrupted faster. Senior professionals clung to legacy systems and mental models.
Generative AI reversed this structure for the first time in technology history.
AI output quality depends on the depth of experience, knowledge, and context that the human brings to the conversation. A senior engineer with 10 years of architecture experience gets fundamentally different output from Claude Code than a junior using the same tool. The same prompt, the same model — but the context gap produces a quality gap that compounds with every interaction.
For the first time, accumulated experience directly amplifies technological advantage. This is a structural singularity.
The Facts Are Brutal
This isn't speculation. The data is already in:
- Software developer employment for ages 22–25 has dropped ~20% from peak (Stanford, 2025)
- Entry-level hiring in AI-exposed roles fell 13% (Stanford, 2025)
- CS graduates now have a 6.1% unemployment rate — higher than philosophy (3.2%) and art history (3.0%) graduates (Federal Reserve Bank of New York, 2025)
- Anthropic's head of Claude Code hasn't written code by hand for over two months — 100% AI-generated (Fortune, January 2026)
- The "10 junior coders → 2 seniors + AI" replacement pattern is already being reported (LA Times, December 2025)
The junior engineer career ladder is collapsing. This is not a future prediction. It is happening now.
The 10:80:10 Rule — A Mental OS, Not a Productivity Hack
Here's what I propose as the foundational framework for human-AI collaboration:
| Phase | What It Means |
|---|---|
| First 10% | Your will. What are you asking? What do you actually want? Without this, you're just drifting on AI output. |
| 80% | AI's output. Let it do what it does best — processing, generating, synthesizing. |
| Last 10% | Your judgment. Is the AI's response aligned with your axis? The moment you surrender this, you become a terminal for someone else's model. |
This is not an efficiency framework. It's a mental operating system for remaining human in the AI era.
Engineers understand this intuitively. Requirements without intent produce technical debt. AI usage without intent produces thinking debt.
Critical Thinking Is Not Academic — It's Self-Defense
When you review a pull request, you ask: "Why this implementation?"
Apply the same discipline to AI output. Ask: "Why this answer? What assumptions is it making? What context is it missing?"
This isn't about being skeptical of AI. It's about maintaining your own axis — your judgment, your values, your professional standards — while leveraging AI's speed.
Critical thinking in the AI era is not an academic luxury. It is a defensive technology.
To Junior Engineers: Arm Yourself
A growing number of young professionals are turning to AI for life advice, career guidance, even emotional support. When you engage AI without your own intent, you don't just outsource thinking — you outsource feeling.
Don't be afraid. But arm yourself.
Learn context engineering. Learn what Andrej Karpathy calls "agentic engineering." But before all of that — have your own axis. Know what you're asking and why. That first 10% is everything. Without it, the remaining 90% is meaningless.
And speak up. No one is going to hand you the practice field. Theory alone doesn't build capability. You need to throw theory against reality, fail, adjust, and loop back. That cycle — theory ⇔ practice — is the only thing that builds real skill.
To Senior Engineers: Honor Your Debt
You are the greatest beneficiary of generative AI. Your 10, 15, 20 years of experience are being amplified like never before.
But are you using that amplification only for yourself?
Think back. Someone reviewed your terrible first PR. Someone explained distributed systems to you on a whiteboard. Someone let you fail on a small project so you could succeed on a big one.
You were raised by the generation before you. Don't break that chain.
Humanity has always evolved by passing knowledge from the experienced to the next generation. The engineering community holds this culture more strongly than any other profession.
AI knowledge — not prompt templates, but the mental OS for thinking with AI — must be part of that transfer.
The Full Book Is Open Source
I wrote an entire book on this topic and published it under CC BY 4.0. Free. No paywall. No signup.
It covers:
- The structural reversal of generational advantage in the AI era
- The collapse of entry-level career ladders (with primary sources)
- The 10:80:10 mental OS framework
- Critical thinking as defensive technology
- A call to action for both generations
📖 **Read the full book:
what-they-wont-teach-you
Top comments (0)