I sat down with Scott Hanselman, Vice President and Member of Technical Staff at Microsoft and GitHub, to talk about what AI coding agents mean for the future of software engineering.
We discussed why AI is not the end of the developer career, what junior software engineers should learn now, and how developers can build real projects that stand out in a changing tech career landscape.
Summary of Lessons
1. Every generation panics about new tools
Scott’s view is that AI anxiety is not new. Developers panicked about color syntax highlighting, autocomplete, Stack Overflow, and now AI. His point: the tool changes, but the craft of software engineering remains.
2. AI is a powerful tool, not a replacement for judgment
AI can generate code, but it does not automatically create good architecture. Developers still need to understand tradeoffs, design systems, review code, and take responsibility for what ships.
3. Treat AI-generated code like code from a stranger
Scott compares AI output to a pull request from an anonymous person on the internet. You should not blindly trust it. Use tests, CI/CD, code review, dependency checks, and a proper software development lifecycle.
4. Human responsibility matters more, not less
If AI creates a bad feature and you merge it, that is still your responsibility. The engineer owns the product, the user experience, and the consequences of the code.
5. AI can create unhealthy productivity dopamine
Scott warns that developers can start “vibrating” with AI enthusiasm. The lesson is to slow down, think clearly, and ask whether you are solving the right problem in a healthy way.
6. Junior engineers still matter
One of his strongest points: the software engineering profession will collapse if companies stop hiring early-career developers. Today’s juniors become tomorrow’s seniors.
7. Senior engineers should be judged by how many seniors they create
Scott argues that mentorship should be formalized. A senior engineer’s role is not just to ship code, but to create more capable engineers.
8. Learn the fundamentals
If he were starting over, Scott would focus on basics: HTTP, DNS, distributed systems, deadlocks, and core software engineering concepts. AI can help you learn faster, but it cannot replace understanding.
9. Communication is part of programming
Programming is not just writing syntax. It is expressing intent clearly so humans, compilers, tests, and AI tools can understand what you mean.
10. Make projects that actually work
Do not just build another tic-tac-toe clone or one-shot a Minecraft clone. Build something useful for your life, family, health, community, church, mosque, hobbies, or interests. Real projects show agency.
A project is more impressive when it is maintained, tested, deployed, documented, and useful.
11. Developers often overcomplicate systems
Scott emphasized simple architecture: static sites, JSON files, caching, GitHub Issues as a database, and avoiding unnecessary infrastructure. The lesson: use the simplest system that solves the problem.
12. Software engineering is still worth learning
His answer to “Will AI replace software engineers?” was: “It better not.” The broader lesson is that the future developer role changes, but the need for people who can understand problems and build reliable systems remains.
Timestamp
0:00 Trailer
1:14 How This Transition Compares To Previous Ones?
2:32 What Percentage Would You Say is AI-assisted In Your Workflow?
4:55 Is Thinking Still On Us?
5:27 What Projects Are You Working On?
7:35 Best Practices for AI Coding?
9:27 How Should Beginners Learn with AI?
12:57 What would you do if you wanted to land a job in 12 months?
14:34 Do Personal Projects Help Land Jobs?
16:12 Do You Have More Demos to Show As Inspiration for People?
20:50 How Do You Manage Your Time?
23:00 Rapid Fire
About Scott
Scott Hanselman is Vice President and Member of Technical Staff at Microsoft, where he works on AI-powered developer tools across Microsoft, GitHub, and Windows. A programmer, architect, teacher, author, and longtime developer advocate, Scott has spent decades helping developers build software and adapt to major technology shifts. In this conversation, we discuss AI coding agents, the future of software engineering, and why he believes there may be more opportunities than ever for the next generation of developers.
More on YouTube related to this episode:
Connect with me
💻 LinkedIn: https://www.linkedin.com/in/jeanklee/
🌸 Instagram: https://www.instagram.com/jeanexplains/
