AI Coding Agents and how to code them
summary
Explains how coding agents actually work by building one live: a REACT loop that issues thought → action → observation and calls concrete tools (ping, bash, web_search) to compensate for LLMs’ lack of local or web access, then wires those tools via MCP (Model Context Protocol) servers (file-system, fetch, Brave search). Demonstrates a Langraph multi-agent setup with multi-agent supervisor→frontend/backend/DevOps producing a React frontend, FastAPI backend, Postgres DB, and shows runnable Python/JS code on GitHub plus an IDE-integrated agent demo. Mentions 70% of dev time is understanding versus 5% editing.
AI-written summary · about
chapters (8)
- 0:00AI coding agents intro
- 7:50Developer time breakdown
- 11:28Basic AI agent with tools
- 17:36Model Context Protocol
- 24:00Multi-agent collaboration
- 31:17VS Code AI assistant demo
- 34:30Future of developers with AI
- 36:20Q&A
related talks
48mAI Coding Agents and How to Code Them
48mAI Coding Agents and how to code them
01h 28mAI Coding Agents and How to Code Them
01h 03mAI Coding Agents and how to code them
01h 28mAI Coding Agents and how to code them
44mHow Does a Coding Agent Work?
27mAgentic Coding with Discipline & Skill: Practical Techniques and Tips
30mBeyond the Hype: Using AI Coding Agents Today