Principles for Building Things That Matter

Long before AI, the most effective teams built great things by moving fast, developing conviction, staying close to users, keeping teams small, and taking responsibility for hard decisions. These principles predate modern management theory, SaaS, and the ZIRP era. They are simply how good builders have always worked.

What is new is the degree to which these principles are now amplified. As coordination and execution collapse into software, judgment becomes the scarce resource. Teams that already know how to move with clarity, ownership, and speed compound their advantage. Teams that rely on process, consensus, and headcount as proxies for progress fall further behind.

The Era of Human Agency does not require new principles. It demands that we finally take the old ones seriously.

• Speed wins. Learn faster. Make decisions faster. Fail faster. Speed is the only competitive advantage in today's world. If we aren't moving, we aren't learning — and if we aren't learning, we will lose.
• Build conviction, not confidence. Be data-informed, not data-driven. As an industry we deeply undervalue expertise and conviction, sacrificed at the altar of analytics and collective decision making.
• Listen to users relentlessly. When you can't find a user, talk to someone who is closest to the user. And then go talk to a user. Be insatiably curious, listen and test with users, build expertise in their problems, and then act decisively. That's 90% of the game.
• Worry about success once you have success. Don't design for problems you don't understand. Plenty of complexity will come if you succeed — so start simple.
• Design your world with the right amount of chaos. Creativity is an uneven motion. If you have no divergent thinking, you will get no creativity. The right amount of productive chaos unlocks the creative output you need to win.
• Absorb complexity so you can broadcast simplicity. The world is glad to be infinitely complex. It is our responsibility to understand that complexity and build ways to simplify it for our users.
• Never hoard information. Always share knowledge. Information is what we observe, track, and store. Information turns into knowledge when we share it. Document everything — from call transcripts to working assumptions — and make it available as collective knowledge.
• Attack problems, not people. Debate hard, hold high standards, and then be humble and commit. Curiosity requires trust, and trust comes from psychological safety.
• Big, persistent problems are the only ones that generate real, sustainable returns. Small and temporal problems are not places to build important businesses, even if solved perfectly.
• Small, lean, focused teams are the only way to solve hard problems well. Quality of people always beats quantity.

These principles assume small teams with real ownership. They reward judgment over activity, conviction over consensus, and learning over comfort. They work because they reduce coordination, increase clarity, and push decisions to the people closest to the problem.

In the Era of Human Agency, these principles have moved from differentiating to required. AI removes the cost excuses that once justified slow decisions, bloated teams, and process-heavy coordination. What remains is the responsibility to choose clearly, act decisively, and own outcomes.

This is not a new way of building. It is a return to what has always driven outsized return, now operating at a much higher frequency and with far less margin for avoidance.

v1.5 — last updated March 2026