Building in the open, slowly

We publish the Lab because most of what we build starts as a question, not a feature. "Would a daily digest beat a sorted inbox?" is not something we can answer from a whiteboard. We have to build the small version and use it.

Some experiments will stay small forever. Some will get folded into the main product. Some will quietly retire. We mark all three states honestly rather than pretending everything in the Lab is on a path to shipping.

This is slower than a private roadmap. It is also more honest, and it keeps us from building things nobody asked to see.