You don't need a small product, you need a small test

Discovery, Lean, and the MVP were built to help you learn whether you're wrong, cheaply. Run as product tasks, they make you efficient at building the wrong thing.

"I ship every week," he told me, like it was the answer to everything.

And he did. Tight build loop. Clean releases. He'd A/B tested his onboarding four times. By every measure of engineering discipline, he was a machine.

He'd also never once tested whether anyone wanted the outcome badly enough to switch off the tool they already had.

So he'd built the wrong thing twelve times, very efficiently.

"Lean" was supposed to help him avoid wasted building. He heard "build fast" and deleted the word "wasted." Speed became the goal instead of learning, and speed without aim just gets you to the wrong place sooner.

A faster plane to a destination nobody wanted is still empty. It just lands sooner.