I'm not asking for perfect, every software has flaws and improves over time and I'm cheering for you. But this is something so fundamental and the fact it's badly broken is not a good look at all, specially when we are talking about something as sensitive as e2e encrypted comms. That's why I asked if you are dogfooding it. Smol demos with fresh npubs don't count. Simulated environments that don't look like real world production usage doesn't count either. I'm talking about using it on a daily basis, using it for all your team's comms until you nail down the onboarding & basics and squash all the obvious bugs, then build from there.

Releasing critical defects that cause 100% breakage for a significant segment of users is far worse than holding back until those defects are fixed. The "release early and release often, even if it's dogshit" has never been a good paradigm. Give weight to defect classifications, keep track of your weighted defect count prior to release, then go into "no new features" mode until the score is low enough AND there are no critical high-priority defects. It's not even challenging to do. It just takes being patient while being excited to release new features. The other option is to have a reputation for releasing garbage, which results in new people not even installing to check it out. Or worse, dedicated users swearing to never install software you make ever again. That's how Norton Utilities went from being the de-facto software for all admins in the PC world, to total bankruptcy.

We deliver shippable intermediaries.