From skeptic to true believer: How OpenClaw changed my life

Claire Vo, host of the How I AI podcast, longtime engineer, three-time chief product officer and founder of ChatPRD, shares how she went from skeptic to power user of OpenClaw, running multiple purpose-built agents for work and family life. The conversation, hosted by Lenny Rachitsky, covers practical setup, security and the mental model of managing agents like employees, plus concrete use cases and common pitfalls.

01Key learnings

• Run several purpose-built agents instead of one general agent; splitting tasks keeps each agent's context focused and reduces forgetting and frustration.
• Install OpenClaw on a separate clean machine such as an old laptop or Mac Mini, with its own local admin account, rather than your main work computer.
• Onboard the agent like a new employee: give it its own email and calendar, share access selectively, and expand trust progressively as it proves reliable.
• Use the best available models for better security against prompt injection and a more reliable experience, rather than cheap models.
• Enable macOS screen sharing and remote login on the agent machine so you can manage it from your laptop without a dedicated monitor and keyboard.
• Use ClaudeCode or Codex on the same machine as a troubleshooting surgeon to read the docs, fix configuration problems and split agent memory.
• Treat agent failures like a manager would: check whether it has the context, tools and clear role definition before blaming the agent, and give it a tools file for how to use each tool.
• Talk to agents conversationally and by voice note, a high-bandwidth way to give rambling context, and have them assign you tasks in your project tracker for real-world to-dos.

02Frameworks mentioned

Summary and takeaways written by PM Atlas; quotes are short excerpts. © the original author.