October 9, 2026
Done by the book
The cargo cults we keep running into, the incentives that produce them, and the slower model we picked instead.
We sometimes look at how other companies in our space operate. Not to copy them. Mostly to understand why they do what they do, because from the outside a lot of it looks like cargo culting. Sometimes it reads like desperation to meet a promise made to investors. It's uncomfortable to watch, and it's not a place I want us to end up.
I'll go through the patterns I keep seeing. We don't have this figured out. These are just the ones we've decided not to follow.
Software got cheap. Getting people to know about you and want to use you didn't. Word of mouth is too slow for investors, unless there's a real network effect, and that is slow and one in a million. So the alternative is to buy attention. Attention is limited, so you go long distances throwing money at psychology. We were the company that gave you the swag. We are the company that helps you make money and live the indie dream you parked in the corner years ago. Some environments turn into a competition for attention that meets developers exactly where they are, which is wanting free stuff. Someone long in business would tell me this is how business works. I'd push back. That is how it has been done, not how it must be done.
Put people in that machine who don't understand the problems of the people they're talking to, and it gets weirder. Especially in our space, where developers want to feel that whoever is talking to them gets their problem. But marketing has to scale, so you hire roles you can send to events to interpret, collect emails, and turn conversations into tickets. I get it. It scales attention capture. It still feels wrong. We put developers in front of developers, and that's the piece we scale. We help them talk about the craft in the context of the problems they're solving and the support they need. People upgrade on their own after that, because they felt heard and because they speak the same language. The language of the tools they use every day.
Product gets weird too. Some products feel stuck in an idea of what they once were. A couple of days ago I saw a Forbes under 30 style post from an investor celebrating a feature that took us an afternoon to build and many months for the industry to see coming. My theory is that not all founders are good. A lot of them got capital onto the right idea at the right time, got elevated through round after round, and ended up believing they own a market. Somewhere along the way they drifted so far from the product that they can no longer steer it. It's like having kids and spending no time with them because you're obsessed with your career, then being surprised to meet a pile of unattended needs and traumas, and trying to fix it with money. I can't think of a better fix than founders building. Tobi at Shopify is the reference there.
It's also easy to give up on the non-obvious. It's easy to jump on what looks obvious, especially once you've said it enough times that people believe it. But if you believe something fills a gap, the first reaction around you will be that it's a bad idea, that no company would justify investing in it. Someone told me the other day that I shouldn't build a build system because Bazel exists. I don't buy it. If you love Bazel, use Bazel, make Bazel better. We improve what we can too, but we think we can do better, especially now, in a very different environment from the one Bazel came from, or the one at Google where most of those decisions happened. That is competition, and it improves things. Buck and Bazel need contenders. That's why we're building Once.
Operations look odd too. Companies are stuck in years of process built to coordinate layers of people and scale throughput. What could go from an idea to a solution in a prompt becomes a ticket in a system the user can't see, waiting weeks for someone to approve it, often a product manager or an executive who wants to feel like they brought something to the table. Weeks later an engineer picks it up. Then someone writes a Forbes under 30 style announcement about what was a prompt. I'd be frustrated working at any of those companies. But I get why they end up there. They run by the book. The book says you scale with people, and the money comes for the people. People need a system, so you build one, juggling emotional creatures who all want to feel important and part of a mission. Naturally it gets slower. Leadership juggles that with investors expecting the next Facebook, and the creativity muscle atrophies.
Craft is the craziest one. I use products and think the people behind them have never used them. Products drift into inconsistency unless someone keeps peeling back the layers so the product evolves as a whole and not as an archaeological site. A button for this customer request, a button for that one. A sandbox renamed to a proprietary name nobody outside the company understands. An upgrade button here, another one there. It's insane, and it's a bad signal about what is possible with software. A company is moving to us right now. Their current tool wants more money to talk to the MCP server of the self-hosted version they already run. The engineer who found out had one reaction. Enough, and that was the end of it. That is what happens when sales and marketing drive the decisions about craft. Financial and attention incentives push you to compromise on craft until the product is a vending machine. And every vending machine is also an attention machine.
I want no part of that. Our model is different. Slower, sure, but it compounds. People don't come because we bought their attention. They come because someone who liked it told them about it. We engineered the organization for efficiency and we're ai-pilled, so we can build good craft fast. Would you go back to a restaurant where you felt they compromised the food? No. So why would you expect your customers to stay if you compromise the craft you put in their hands? Money can buy you the first dinner. It cannot buy you a customer who comes back. Only the language of good craft does that. It isn't written in the books that sell you GTM tricks, but care for craft goes a long way in business. If you don't care, no one cares.