We Studied 150 Developers Using AI (Here’s What's Actually Changed...)

AI coding tools like Claude Code, GitHub Copilot, Cursor, and ChatGPT are changing how we write software — but are they actually improving long-term software quality, or just accelerating slop?

In this episode of Modern Software Engineering, Dave Farley explores the real impact of coding with AI, drawing on published research into human vs AI code maintainability. We look at what happens after the AI has written the code — when the next developer needs to understand, change, refactor, and maintain it.

You’ll learn:

Whether AI programming actually harms or helps maintainability
Practical Claude Code tips and insights into how Claude Code agents influence developer productivity
The truth behind vibe coding — including what is vibe coding and where it helps (or hurts)
Why AI vs human debates miss the real issue: engineering skill and design thinking
How AI can amplify good practices — or accelerate bad ones
What serious software engineers should focus on in the age of artificial intelligence

This video goes beyond how to use Claude Code and dives into what actually matters in software development: maintainability, design quality, decomposition, and long-term engineering discipline.