What I've Learned
Practical insights from building products, scaling teams, and embracing AI
Human in the Loop: Why AI Still Needs Human Oversight
AI writes code faster than any team, but it can't tell when it's wrong. What human in the loop means, what breaks without it, and how to check your team has it.
Travis Sutphin
Principal engineer. 25+ years building software.
Prompt Engineering vs Prompt Guessing: When to Use Each
Prompt engineering vs prompt guessing: why prompts without context, guidelines and goals give random results, and when guessing is actually the right move.
Travis Sutphin
Principal engineer. 25+ years building software.
When to Use AI Thinking Mode (and When to Turn It Off)
When to use AI thinking mode: turn it on for strategy, tradeoffs and open questions, and off when you know exactly what you want built. A simple rule of thumb.
Travis Sutphin
Principal engineer. 25+ years building software.
Vibe Coding for Beginners: 5 Do's and 5 Don'ts
A vibe coding guide for beginners: 5 do's and 5 don'ts for building with AI tools, from small first projects and clear prompts to testing and version control.
Travis Sutphin
Principal engineer. 25+ years building software.
Vibe Coding with AI: Why Planning Still Matters
Vibe coding with AI gets you to 80% fast, then stalls. Why planning still matters, and how having AI document its own work keeps your codebase consistent.
Travis Sutphin
Principal engineer. 25+ years building software.
AI Development Best Practices: Avoid the AI Speed Trap
AI doesn't fix chaos, it amplifies it. AI development best practices for moving fast with systems: documentation, tests, prompt templates and daily reviews.
Travis Sutphin
Principal engineer. 25+ years building software.
AI Project Stuck at 80%? How to Finish and Ship It
Why AI projects get stuck at 80%: the five blockers behind the last 20%, how to tell if your project is really stuck, and two practical paths to ship it.
Travis Sutphin
Principal engineer. 25+ years building software.