Jacob Spencer

Following up on the first post in this series and the thread it kicked off on app distribution. I ran a two-week observation on organic distribution and want to bring the results back for discussion, not conclusions. Translation tool was used again, so please read past awkward phrasing.

The reason I am posting again : the same question that came up last time is still open for me, and I now have two weeks of data that only made it sharper. I would rather test my reading against yours than sit on it.

When was the last time you built a feature because one user expressed a strong frustration?

It happens easily. A detailed complaint feels urgent, the solution seems obvious, and suddenly the feature is on the roadmap.

But a user asking for something does not always mean they would use it, pay for it, or even notice when it ships.

Before adding a feature to the roadmap, I find these five questions useful:

What breaks when your AI agent needs a human answer?

I have been seeing the same pattern across coding agents and desktop agents: the agent can do more of the work now, but the flow still depends on small human approvals at the worst possible moments.

A shell command needs permission. A file change needs review. A calendar action feels too risky to run automatically. A tool call looks safe in one repo but dangerous in another. The agent pauses, the user is away, and the whole run loses momentum.