AI Coding Agents and How to Code Them
summary
Demonstrates building coding agents that use React (reason+act) framework with concrete tools: ping, bash and web search hooked into GPT-4.1, then swapping to Entropic MCP servers via MCP (Model Context Protocol) to expose filesystem, fetch, and web‑search tools. Uses LangChain/l‑graph to compose a supervisor plus front‑end, back‑end and devops agents that generate ~20–30 files (TypeScript React frontend, FastAPI backend, Postgres, Docker Compose) and actually spin up a site, though one backend endpoint fails; cites that understanding takes ~70% of developer time versus 5% for editing and shows encoder VS Code/JetBrains plugin enabling the same agent flows.
AI-written summary · about
chapters (12)
- 0:00Introduction and agenda
- 3:15What is an AI agent?
- 7:27Why agents matter to developers
- 10:58Baseline LLM setup
- 14:10Building a basic agent
- 21:23Introducing MCP protocol
- 25:44Creating MCP servers
- 31:50Multi-agent collaboration
- 37:53Using existing agent tools
- 41:24Live demo results
- 43:11Future of developers with AI
- 46:24Q&A and closing
related talks
40mAI Coding Agents and how to code them
48mAI 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
01h 28mAI Coding Agents and how to code them
30mBeyond the Hype: Using AI Coding Agents Today
01h 01mAI Agents In-Depth – Function Calling, MCP and Tool Use Under the Hood
27mAgentic Coding with Discipline & Skill: Practical Techniques and Tips