We used to run a lot of "alignment meetings." Lay out directions, argue for an hour, pick one, go build it. Felt productive. We were making decisions.
Then we started counting the cost. That hour was our sharpest hour. We were spending it picking which idea to pursue before we had any real information about which one was better. Building has gotten so fast with AI tooling that the meeting itself became the bottleneck.
So we changed the rule: everyone takes the same brief and builds their own version all the way to a working product. Then we look at what exists and decide.
The first time we ran it, three of us took the same one-paragraph brief and disappeared for a day. No wireframes, no check-ins. Each person just built.
A day later: three working products in three browser tabs.
Two solved the problem the way we'd have converged on in the meeting. The third solved it in a way none of us would have proposed out loud. Different entry point, different mental model, a structural decision that looked wrong until you used it. That third version is the one that shipped.
The meeting we skipped would have killed it in its first sentence. It's an idea that doesn't make sense until you see it run. You can't argue for it. You can only build it.
The obvious objection: four people building four full products is chaos. It is, without the right setup. We built DevKanban for exactly this — each direction on its own card, its own branch, its own cloud environment, none touching each other. When it's time to decide, you're not choosing between descriptions. You're choosing between things that run.
The pitch meeting selects for ideas that are easy to describe. Not ideas that reveal themselves through use. Those die before anyone builds them.
When you build first, you find out which ideas actually work. The meeting still happens, it just takes fifteen minutes instead of an hour.
Do you debate before building, or build first? Curious what made the switch click for other small teams.
We're Team Sisyphus, a small team building AI tools for AI native builders and teams.
I like the shift from debating possibilities to comparing evidence. When building gets cheap, the scarce resource isn't engineering time anymore—it's confidence that you're backing the right idea. Working prototypes generate a kind of feedback that conversations rarely can, especially for ideas that only make sense once someone experiences them.
yes, exactly.
it could also become the competitive edge in terms of speed vs. quality of product in the near future
I think that's where the more interesting decision starts.
Speed and quality can both improve.
The harder question is how you decide that the evidence from a prototype is actually strong enough to justify committing the business in one direction.
I've got a thought on that which probably doesn't fit well in a thread. If you're interested, what's the best email to reach you on?