You are the main thread: Multiplying your productivity with agents
summary
Practical guide to running parallel dev agents: use Claude or Cursor via Cloud Code, write an agents.mmd file with build and formatter rules, and connect MCPs like the Sentry MCP to read crashes and create PRs. Demonstrated wins: auto-updating Jira/Linear tickets, addressing PR comments, rebasing and resolving merge conflicts, and porting a duplicate-file analysis from TypeScript to Python; CI fixes work about half the time. Run background workers with git worktree to run multiple agents and always double-check generated code and clear context frequently.
AI-written summary · about
chapters (20)
- 0:00AI productivity with agents
- 1:10Lazy developer philosophy
- 3:55From Android dev to AI backend
- 5:40AI tools convergence and adaptation
- 6:22Maximizing tokens with Ultra Think
- 8:50Backend clients and frontend code
- 10:16Rewriting code with AI references
- 11:19Automating testing workflows
- 13:29Executing tests and CI checks
- 14:09Writing commits and PRs
- 15:32Slash commands for Git
- 17:37Diagnosing CI failures
- 19:45Addressing PR comments
- 20:28Model Context Protocol
- 21:51Planning and resolving issues
- 23:58Debugging features and production
- 25:43Managing context windows
- 27:28Running multiple agents
- 29:14AI-driven development lifecycle
- 30:18Q&A on AI agents
related talks
40mHow To Tame Your Agents?
01h 03m10 tips to level up your ai-assisted coding
01h 01m10 tips to level up your ai-assisted coding
25mA Unified Agent Experience
27mAgentic Coding with Discipline & Skill: Practical Techniques and Tips
39mAgents for the Sake of Happiness
30mBuilding Tools with Tools: Modern AI-Driven Development
17mCooking with Agents in VS Code