AI Coding Tools: Hype vs. Reality
A frank assessment of AI coding assistants after 12 months of daily use. What works, what's oversold, and what leaders should actually expect.
Show Notes
After a year of daily use, I give my honest assessment of AI coding assistants and what they mean for engineering teams.
Key Points
- What genuinely works — boilerplate elimination, code explanation, test generation
- What’s oversold — architecture decisions, novel problem-solving, the idea that AI replaces developers
- What leaders need to know — these are productivity tools for skilled engineers, not replacements for engineering judgement
Links Mentioned
- Blog post: “AI Coding Assistants: An Honest Review After 12 Months”
- GitHub Copilot productivity study
- Stack Overflow Developer Survey 2025
Timestamps
- 00:00 — Introduction
- 03:15 — The tools I tested
- 08:00 — What actually works well
- 18:30 — What doesn’t work (and why)
- 26:00 — Advice for engineering leaders
- 33:00 — Wrap-up