#Values

What we optimize for as an engineering org — the handful of principles that settle the argument when two good options compete. Kept short on purpose. If a value doesn't change a decision you'd otherwise make, it's decoration; cut it.

#Create value, not code

The best code is the code we never wrote. We're paid to solve problems, not to ship lines — so we buy before we build, integrate before we invent, and delete before we add. Every line we write is a liability someone maintains forever; the win is the outcome in front of a user, not the size of the diff. Spend our real engineering only on the problems that are actually ours to solve.

#Default to no

Most good ideas still get a no, and that's the point. Focus is subtraction — every yes is a forever cost in surface area, maintenance, and attention, so the bar to add is high and the bar to decline is low. "Why does this even need to exist" is the first question, not the last. We'd rather do five things at a level nobody else can touch than twenty things merely well.

#Be opinionated

There's usually one good way to do a thing, and it's our job to have a view on what it is. Strong defaults over endless flexibility; taste over "let the user decide"; conviction over a neutral menu of options. We don't A/B our way to a personality. Bring the opinion and argue it hard — then change your mind the instant someone's is better. Strong views, held loosely.

#Obsess over craft

Quality is the first principle — everything else flows from it, and we don't trade it for a deadline. Craft isn't a phase at the end; it's how we work the whole way through, down to the names, the edges, the error paths, and the thing that breaks at 2am. Fast-and-sloppy isn't fast; it's debt with interest, and it always comes due. If you wouldn't put your name on it, it doesn't ship.

#Own it end to end

Every project and every issue has one accountable owner — you build it, you run it. Ownership runs from the requirement to the pager: the review, the deploy, the docs, the on-call, the outcome in production. No throwing it over a wall and calling it done. If you can't operate it, you haven't finished building it.

#Momentum over motion

Ship the smallest correct thing, then keep shipping. Progress compounds; big-bang launches rot in a branch while the world moves on. Bias to delivery, put it behind a guardrail, and let production teach you what a planning doc can't. Motion is being busy. Momentum is closing the distance to done.

---

The bar we're aiming at: The Linear Method.