Prefer Feedback to Planning

We believe that great work comes from testing hypotheses and that gathering data is the best way to validate and iterate on ideas. Faced with uncertainty, we trust ourselves to act responsibly.[1] We try our best to embrace the mindset of "no blockers" and non-perfectionism.[2] We believe that measuring work is the first step in improving it and that the earlier we can collect signal, the better.

There are caveats: consequential and irreversible decisions require forethought.[3] In a letter to shareholders, Jeff Bezos described decisions as "one-way doors" and "two-way doors" and cautioned against one-size-fits-all thinking.[4] But when in doubt, we err on the side of action.

Do One Thing and Do It Well

We draw inspiration from the first principle of Unix philosophy: "Make each program do one thing well."[5][6] We believe that complex problems can be broken down and solved with simple and modular components.[7] Programming is no different in this regard from writing.[8]

This principle extends to how we work together. Just as a program should have a clear, focused purpose, so should we. Consider a computer processing tasks: I/O-bound tasks can be efficiently parallelized, but CPU-bound tasks require dedicated focus. Deep, meaningful work is like a CPU-bound task: it demands our undivided attention.

Invest in Leverage

We aim to invest in processes that lift our users and our team.[9] This includes processes like documentation and automation. We believe in explicit communication over hidden state: we strive to approach each other with curiosity and to encapsulate our learnings so that we may all benefit. By investing in tools and automation, we multiply the impact of each of our work. When we build systems that make each other more productive, we create compounding benefits.

Be Pragmatic

We have ambitious goals, but we believe in the power of small, approachable, and consistent steps. We strive to temper idealism and be willing to work with existing systems. Even these values are likely to evolve as the organization does. We have written down the ones we believe will best stand the test of time.[10]