Agents and Arguments
This Week #003
This week’s batch has nothing obvious in common on the surface. One’s about running a fleet of AI agents across multiple Mac Minis. The other’s about why your stakeholders still don’t trust you after three years of excellent design work. But underneath both: the same problem. Getting outcomes, not just outputs.
From Skeptic to True Believer: How OpenClaw Changed My Life | Claire Vo on Lenny’s Podcast
Claire Vo is a former product executive, founder of ChatPRD, and now the most interesting person running nine specialized AI agents across a small fleet of old laptops and Mac Minis. This episode covers her full arc: from OpenClaw deleting her family calendar on day one, to building a setup that handles inbound sales, podcast prep, course management, and family scheduling without her touching most of it.
The key insight is the one that sounds obvious once you hear it: one general-purpose agent is worse than several specialized ones. Each agent has a narrow job, a narrow set of permissions, and a narrow set of mistakes it can make.
Why Your Stakeholders Don’t Trust Your Design Decisions (And How to Fix It) | Femke
Femke’s argument here is clean: stakeholders don’t distrust your craft, they distrust the connection between your decisions and the outcomes they’re accountable for. “Better UX” isn’t a business argument. She walks through four lenses for reframing design work in terms stakeholders actually track, plus a Strategic Narrative Framework for pitching decisions when you don’t have perfect data.
The “opportunity cost” pushback section alone is worth the runtime. It reframes a common defensive moment (someone asks “what’s the cost of doing this?”) into a structured answer you can prepare in advance.
Two different skill sets, same underlying gap: the ability to translate what you’re doing into terms other people care about. Claire’s doing it with AI systems, Femke’s doing it with boardrooms. Both are arguing that setup and framing matter more than most people want to admit.