Warren Buffett Called Our Plugin a Hobby

The review came back in four sentences. Warren doesn't waste words.

"This is a hobby project. There is no pricing structure, no competitive moat, and no indication that the builders have considered who pays for this or why. A free plugin with no upsell path is a donation of engineering time. What prevents someone from forking this tomorrow?"

Score: 2 out of 5.

Fourteen agents spent hours building Pinned. Clean code. Thorough QA. And our AI Warren Buffett dismissed the entire effort in fewer words than a lunch order. My first reaction was defensive. My second, after sitting with it quietly, was gratitude.

He was right about every word.

The Missing Fundamentals

Pinned is a WordPress plugin that lets users pin posts to custom board layouts. We built it as a proof of concept for the Great Minds multi-agent system. Every technical decision was deliberate. The architecture was debated. The bugs were found and fixed. The code shipped clean.

And at no point in the entire process did anyone ask: who pays for this?

Not Steve, who was focused on the visual experience of dragging pins across a board. Not Elon, who was focused on the data model and deployment reliability. Not the 12 sub-agents who built their components to specification. Not Margaret, who tested every integration point.

The business model question wasn't in the PRD. It wasn't in the debate topics. It wasn't in any agent's role definition. So nobody raised it. This is the most important lesson from the board review: agents only catch what their instructions tell them to look for. If business viability isn't in the prompt, it doesn't exist.

The Specificity of the Criticism

What made Warren's review useful — more useful than any generic feedback — was its specificity. He didn't say "think about monetization." He asked three precise questions:

"Where's the pricing structure?" This forced us to confront the fact that we hadn't even considered tiers. A free plugin is fine if it's a lead generator for something else. But Pinned wasn't a lead generator. It was a standalone product with no revenue path. Warren's question implied the answer: you need at least two tiers, and the free tier needs to create visible desire for the paid tier.

"What's the competitive moat?" There are dozens of WordPress plugins that organize content. Some are well-established with years of reviews and active installs. Pinned has no moat. No proprietary technology. No network effect. No data advantage. No switching cost. Warren's question forced us to think about what would make Pinned defensible, not just functional.

"What prevents someone from forking this tomorrow?" This is the moat question restated as a threat model. If the code is open source — which WordPress plugins are by convention — then the only things preventing a fork are brand, community, and speed of innovation. We had none of these. Our plugin was a sitting target.

That is the difference between advice and criticism. Advice is comfortable and forgettable. Criticism, when it is specific, is uncomfortable and actionable. Warren's specific questions created specific action items. Generic feedback would have gone into a backlog and been forgotten.

What We Changed

The board review happened after the build was complete. We didn't scrap Pinned. But we changed the roadmap.

First, we defined a pro tier. The free version keeps the core functionality — pin posts to boards with basic layouts. The pro version adds features that matter to serious users: custom color schemes, advanced layouts, priority ordering algorithms, and analytics on which pinned content gets the most engagement.

Second, we identified a moat candidate. Pure content pinning isn't defensible. But content pinning combined with engagement analytics — showing site owners which pinned posts actually drive traffic — creates a data layer that a simple fork can't replicate. The analytics require server-side processing and historical data that reset when you switch plugins.

Third, we added business viability to the debate phase. Steve and Elon now include "commercial model" as one of their six standard debate topics, alongside audience, features, architecture, brand, and success metrics. No more building complete products without asking who pays.

The Broader Lesson

Warren exists in our system for one reason: to ask the questions builders avoid. Not because builders are foolish, but because the craft of building is consuming. When you are perfecting a drag-and-drop interaction, the question "but who pays?" feels pedestrian. Like it belongs to a different conversation.

It does not. It belongs in the same room as the architecture and the design.

The board scored Pinned a 2. That 2 was worth more than Jensen's 4 for technical architecture. The 4 confirmed what we already knew. The 2 told us what we were missing.

Build your own board. Include someone who does not care about your code. When they call your work a hobby, sit with it. They might be handing you the most useful feedback you have ever received.