Vibe Coding , MCP and AI Agents - Rizel Scarlett

In this conversation, Rizel Scarlett joins me and dives into how AI is reshaping the way developers work. Rizel walks through what Goose actually is, why MCP matters, and how agents move beyond autocomplete into real autonomy. We get into practical workflows, how to keep agents from hallucinating, what tasks you should hand off vs keep for yourself, and why the Chrome DevTools MCP server changes the game. If you’ve been curious about agentic workflows, vibe-coding, or the future of engineering productivity, this episode lays it out in a way that actually makes sense. Timestamp: 00:00 — introduction 02:45 — What exactly is Goose? 05:10 — How does Goose differ from a normal LLM? 07:58 — Can AI help developers solve harder problems? 09:40 — When should you trust an agent vs doing it yourself? 11:12 — What does a healthy AI-powered workflow look like? 14:40 — What are MCPs and why do they matter? 17:30 — How does the Chrome DevTools MCP server help engineers? 19:55 — What does “vibe coding” actually mean? 23:30 — How do you avoid hallucinations from agents? 25:20 — What are the best ways to use AI in development today? 27:05 — Which tools or workflows are worth learning now? 29:10 — Where is AI tooling heading next?