How we work · 4 min read
Meetings and tools
How we run the few meetings we keep, why we default to async, and the tools we reach for.
We are not a meetings company. Every hour a group of people spends in a call is an hour nobody spends building, and the math on that gets brutal fast when you multiply it out. So we keep meetings rare, short, and worth the interruption. Most of what other companies would call a meeting, we handle in writing instead. This page is about the exceptions and the tools that let us live that way.
Agenda or it's cancelled
If a meeting doesn't have a written agenda before it starts, it doesn't start. That's not a bureaucratic hoop, it's a filter. A meeting nobody could be bothered to write three bullet points for is a meeting nobody needs to attend.
The agenda is a shared document, not a slide deck. You write down what you want to decide or discuss, people read it ahead of time, and the meeting is for the part that actually needs voices in a room: the disagreement, the judgment call, the thing that's faster to talk through than to type. If a topic can be resolved by everyone reading and commenting, it never becomes a meeting at all.
Notes always
Every meeting produces notes, and the notes outlive the meeting. Decisions, owners, and open questions go in the doc while it's fresh, and the doc goes somewhere the whole company can find it. This is what makes it safe to not invite everyone. Nobody has to sit in a call defensively just to stay informed, because the record is public by default and they can catch up in five minutes. A decision only the attendees remember is a rumor, not a decision.
Default to async
We're small and increasingly spread out, and we treat that as a feature. Async is the default because it respects focus and it respects time zones. You write the thing up, people respond when their day allows, and the good version of the idea wins instead of the loudest voice in the fastest reply. This ties directly into how we handle communication: clear writing is the core skill, and a meeting is what you reach for only when writing has genuinely stalled.
Synchronous time is precious, so we spend it on the things it's actually good at: building trust, untangling a real disagreement, celebrating something, or the occasional call where you just need to hear a human. Not status updates. Status belongs in a doc.
Documents over decks
We write documents, not presentations. A deck is optimized for the person presenting; a document is optimized for the person reading, and it forces an actual argument in full sentences instead of six bullets and a chart you talk over. If your thinking survives being written out, it's probably sound; if it only works with you narrating, it wasn't ready. How we write goes deeper.
Calendar hygiene
Your calendar is a shared resource, so keep it honest. Block the time you need to do deep work and defend it. Decline meetings you don't need to be in; a polite "I don't think I'm needed here, but send me the notes" is a gift, not an insult. Keep your working hours and time zone visible so nobody has to guess. If a recurring meeting has quietly stopped earning its slot, kill it. Recurring meetings should have to justify their continued existence, not coast on inertia.
The tools we reach for
We describe our tools by the job they do, not the brand on them, because brands change and the jobs don't. We keep the set small: fewer tools means fewer places to look and fewer accounts to wrangle.
- Where the code and its history live. Everything of consequence goes through version control and code review. The commit log is our real system of record.
- Where we talk. A single chat space for the fast, low-stakes back-and-forth, with public channels by default so conversations are searchable later.
- Where we think out loud. A shared writing space for docs, proposals, and meeting notes, one that makes commenting easy.
- Where the work is tracked. One lightweight issue tracker, close to the code, so what's planned and what's in flight are visible without a meeting.
The rule underneath all of it: if a tool starts creating more overhead than it removes, we drop it. The tools serve the work, never the other way around.