Building in Public: Demox's First Month

Demox has been live for about a month. I said from day one I'd build this in public, so here's the honest update — the numbers, the wins, and the stuff that isn't working yet.

The numbers. We have roughly 50 registered users. About 100 posts. Four active communities. A few dozen comments a day on good days, single digits on quiet ones. If you're coming from a world where "successful launch" means 50,000 signups in week one, these numbers look small. They are small. They're also real humans, and I can tell you that because I've read most of what's been posted.

I'd rather have 50 people who actually talk to each other than 50,000 accounts that signed up, looked around, and never came back.

What's working. The AI moderation has handled every flagged item without producing a single appeal I'd call embarrassing. The rules are narrow, the decisions line up with the rules, and the modlog is public so you can check my work. That was the scariest part of the design — whether AI moderation would feel arbitrary in practice — and so far it feels consistent.

The "no email signup" flow is doing exactly what I hoped. People sign up in under 30 seconds. A handful of users told me they wouldn't have bothered if I'd asked for an email. That matches my bet: friction at the door kills a platform before it starts.

The news aggregation keeps the communities from looking empty. When someone visits a quiet community, there's fresh content to react to. It's not a replacement for organic discussion, but it's a bridge.

What's broken. Mobile is rough. The site works on phones, but "works" is doing heavy lifting in that sentence. Some interactions are awkward. The vote buttons are too small on narrow screens. The comment threading gets cramped. I know. It's next on the list.

Search is weak. Right now it does a basic text match and not much more. If you're trying to find an old post, good luck. Proper search is harder than it looks, and I haven't prioritized it yet because the content volume is small enough that browsing works. That calculation changes as we grow.

Community discovery is thin. We have four communities. If you don't know they exist, you might not find them. I need a better way to surface them on the landing page without turning the landing page into a billboard.

Notifications are basic. You get one when someone replies to you, and that's about it. No digest, no batching, no preferences. Fine for now. Not fine forever.

What's next. More communities — I'm taking suggestions, so if there's a topic you want a home for, tell me. Better mobile, because the current state is embarrassing and I want to stop being embarrassed by it. A real search. And I want to start letting the community suggest features directly, voted on transparently, instead of me deciding everything in a vacuum.

Is a month enough to know if this works? No. Ask me again at three months, and again at a year. I'll keep posting the numbers.

Thanks for being here early. It matters.

— Tom