A five day series on building real software with agentic AI, one post per day.
Why AI only gets you 80% of the way, how to stand out in a flood of mediocre software, and how to build real leverage.
Day 2The actual tools you need to know to get high quality output from AI instead of rookie results.
Day 3Surrounding nondeterministic AI with deterministic checks: strict TypeScript, custom lint rules, stop hooks, and review subagents.
Day 4Giving AI as much freedom as possible without the risk: permission modes, surgical allowlists, sandboxing, MCP, and pre-command hooks.
Day 5A systematic approach to using subagents to build larger features and improve existing software.
An end-to-end program on utilizing AI the right way. 10X your productivity, build self-verifying AI (no more AI mistakes), and have autonomous AI run for hours.