Dock Dock All notes

Why we're starting again

In September 2025, a nonprofit that teaches teenagers to code was given seven days to find $50,000.

Hack Club had been on Slack for eleven years, paying $5,000 a year under a nonprofit agreement. Then the number changed. Slack told them it was $50,000 up front and $200,000 a year after that, and that if they didn't pay, the workspace would be switched off and eleven years of messages deleted.

They wrote about it publicly. It went to the front page of Hacker News, The Register picked it up, and Slack's CEO got in touch personally to put it right. The bill went away.

Hack Club moved to Mattermost anyway.

That last part is the bit I keep coming back to.

The quieter version

Almost nobody gets a $200,000 invoice. Plenty of small teams get the smaller version of the same feeling. The per-seat bill that grows every time you hire someone. The free plan that quietly stopped showing anything older than ninety days. The slow realisation that the conversation your company actually runs on lives somewhere you have no claim to, priced by someone who knows how hard it would be to leave.

We had been hearing this for about a year, mostly from small businesses. Rarely as a crisis. Usually as a shrug.

What we found when we went looking

We assumed somebody had already solved this and just marketed it badly.

They hadn't. There is a long tail of team chat products and very few of them clear the bar. Some are capable and genuinely unpleasant to use, the kind of interface that makes sending a quick message feel like filing a ticket. Some are self-hosted only, which is a fine answer if you employ someone to run servers and no answer at all if you're nine people. Some are slow enough that you feel every keystroke. Some get vaguer about where your data actually sits the closer you read. A few cost about as much as the thing they'd be replacing.

What we couldn't find was the plain combination: as good to use as Slack, priced for a company that counts seats carefully, with the data somewhere you can point at on a map.

That looked like a real gap. So we built for it.

What the first version was

It was meant to be a foundation rather than a pitch.

We shipped it to a waitlist of 400 people who had signed up because they said they wanted exactly this.

What happened

Not a rejection. A rejection would have been easier to read.

People signed up. A good number installed it. Some invited their team, sent messages, told us it felt good, and meant it. Then, mostly, they went back to Slack.

We spent a while treating that as a product problem, which is the flattering reading, because product problems have fixes. You can ship your way out of a product problem.

It wasn't one, or it wasn't mainly one. That deserves its own note and it's the next one: People complain, but they don't switch.

What starting again means

Hack Club moved. They moved because the alternative was losing eleven years of their own history inside a week.

That's the shape of it. Teams don't move their conversation because a better one exists. They move when staying stops being possible. Cheaper, faster and better behaved is not a reason to switch. It's a reason to agree with us, and we have 400 people's worth of agreement to prove those are not the same thing.

So we're going back to the foundation. Same subject, how teams work together, and a different answer to it. We're keeping what was never the problem: the speed, the local-first architecture, the position that your data is yours and not leverage. We're throwing out the assumption underneath all of it, which was that a better version of something people already have counts as a product.

What replaces that, I don't know yet. That is honestly where we are. This site exists so we can write it down as we go, including the parts that don't make us look good.