The Blog
Lessons from shipping web and mobile products — engineering, design, product, and the business side of building software.
Most six-week MVPs fail because the team tries to build a three-month product in six weeks. The trick isn't working faster — it's building less. Here's the sequencing we use.
Rewriting a legacy system is risky — but rewriting in-place, route-by-route, alongside the old one, lets you ship the new stack while the business keeps running. Here's how we've done it five times.
Linear, Superhuman, and Bloomberg look dense on purpose. Airbnb, Notion, and Stripe look roomy on purpose. Here's the rule we use to pick a density for a given product.