Introduction
Imagine being the youngest instructor at Stanford University, teaching C++ to some of the brightest minds, while also organizing one of the biggest hackathons in the world. That's the reality for Rachel Fernandez, a computer science student who grew up in Westminster, California—a small town where 70% of her high school peers qualified for free lunches. She was the first student from her school to attend Stanford in years. In a recent podcast interview with Quincy Larson, Rachel shared her hard-earned insights on computer science education in 2026, the enduring relevance of C++, and—most critically—how developers can use AI tools without losing their core skills. This guide distills her advice into actionable steps, so you can leverage AI to amplify your growth rather than replace it.
What You Need
- A computer with internet access
- Basic familiarity with programming (any language)
- Willingness to learn from both success and failure
- Access to AI coding assistants (e.g., ChatGPT, GitHub Copilot, etc.)
- Patience to practice foundational concepts without shortcuts
- A notebook or digital document for reflection
Step-by-Step Guide to Using AI Without Deskilling
Step 1: Ground Yourself in a Foundational Language Like C++
Before you lean on AI, build a solid foundation. Rachel, who teaches C++ at Stanford, emphasizes that low-level languages force you to understand memory management, pointers, and how computers actually execute code. This understanding makes you a better developer even when using high-level abstractions. Spend at least three months writing C++ from scratch—no autocomplete, no copilots. Write a simple database, a file parser, or a game. Once you grasp the basics, you'll recognize when AI-generated code is efficient versus bloated.
Step 2: Participate in Collaborative, Time-Boxed Projects
Rachel helped organize TreeHacks, Stanford's annual hackathon that narrowed 15,000 applicants to 1,000 participants for a weekend of building and competing for $1 million in prizes. Hackathons force you to code under pressure, make quick decisions, and work in teams. They expose gaps in your knowledge that AI can't fill if you don't know what questions to ask. Join a local or virtual hackathon without using AI for the first hour—ideate, sketch architecture, and write pseudocode by hand. Then bring in AI tools only to accelerate known patterns, not to generate novel logic you don't understand.
Step 3: Treat AI as a Co-Pilot, Not a Replacement
Rachel’s key warning: "Don’t deskill yourself." Use AI to check your work, explore alternative solutions, or handle boilerplate code. But never accept suggestions blindly. For every AI-generated block of code, ask yourself: Can I explain every line? If not, rewrite it manually until you can. A good practice is to write a function yourself first, then ask AI for a more efficient version and compare the two. Note the differences in your journal. Over time, you'll internalize better patterns without losing your own problem-solving muscles.
Step 4: Embrace Rigorous Data Quality and Validation
Bad data is the silent killer of AI-assisted projects. Rachel’s experience teaching and hacking has shown her that even the smartest model fails if fed garbage. Use the freeCodeCamp Data Quality Handbook to learn validation layers: frontend, backend, database, business logic, and ingestion. Implement unit tests that catch bad data before it reaches your AI tools. This habit keeps your reasoning sharp and your outputs reliable.
Step 5: Build Responsible AI Systems with Governance
AI governance isn't just for managers—developers build it. Rachel advocates for hands-on projects like those in the freeCodeCamp AI Governance Handbook. Practice building a model card generator, a bias detection pipeline, an audit trail logger, and a human-in-the-loop system. These projects force you to think about ethics, transparency, and accountability. They also make you a more thoughtful engineer who can spot when an AI tool is producing biased or harmful output.
Step 6: Automate Your Routine Tasks, Not Your Learning
Rachel suggests using automation to free up mental bandwidth for complex problem-solving. The freeCodeCamp Automation Course teaches you to connect triggers and actions using Model Context Protocol (MCP). Build servers that share info between your productivity apps and agents. But resist the temptation to automate learning: always write code for new concepts manually until you grok them. Use automation after you understand the underlying process.
Tips
- Reflect daily: Spend 10 minutes journaling what you learned vs. what an AI did for you. This builds metacognition.
- Teach someone else: Like Rachel teaching C++, explaining concepts solidifies your understanding.
- Limit AI usage on difficult problems: Try solving them entirely on paper or in a plain text editor first.
- Leverage community discussions: The freeCodeCamp book on AI-Assisted Coding offers strategies, but use it as a supplement, not a crutch.
- Stay curious: Rachel’s journey from a small town to Stanford shows that curiosity and persistence beat raw talent. Always ask "why?"
For more inspiration, listen to the full podcast on the freeCodeCamp.org YouTube channel or your favorite podcast app. And remember: AI is a tool, not a teacher. The learning happens when you put the tool down and think for yourself.