The best engineers are not fast because they are great. They are great because they are fast. Speed is not the reward for skill. It is the thing that builds the skill, because moving quickly puts you in front of reality more often, and reality is the only teacher that tells you anything true.
Activity alone is just motion. What matters is the loop: produce information, learn from it, produce again. You learn the wall’s thickness by going through it. Put a wall in front of you and there should be a you-shaped hole in it, because standing in front of it, sketching how thick it might be, teaches you nothing. The hole returns a reading. The reading changes the next swing.
This is the line between speed and haste. Haste keeps walking into the same wall. Speed walks through, reads what it cost, and aims the next one better. Fast is how quickly you close the loop and open the next. Loop rate is the metric. Everything else is vanity.
Watch what gets built when the loop is tight. Git was usable on its first day and self-hosting within the week. Brendan Eich wrote the first JavaScript in ten days. Kelly Johnson’s team flew the P-80 prototype 143 days after starting it. Apollo 8 went from a summer decision to a Christmas orbit of the moon in about four months. These were not minor hacks. They were consequential things built under pressure, and the pressure is part of why they were good.
So slow is not the safe choice people think. Slow is risky, because the world moves while you deliberate and your information goes stale. Slow is immature. Slow is average, because it is what everyone defaults to. Slow is a distraction, because it fills the time that contact with reality was supposed to fill.
A redo used to cost months and money, so holding for certainty made sense. Now the cycle is almost free. AI did not make judgment unnecessary. It made the cost of testing judgment collapse. And when the cost of being wrong falls far enough, the whole logic of caution inverts.
The old playbook assumed each wrong turn was expensive, so ambition had to enter through a wedge. Start small. Learn slowly. Earn the right to expand. That was a strategy built for costly loops. When the cost of testing collapses, the wedge stops looking prudent and starts looking mispriced. You can run a thousand cycles in the time the wedge took to find, so the rational move is no longer the cautious entry. It is the whole thing, from the start, faster than looks responsible.
Bezos had the business version: a ten percent chance at a hundred-times payoff is a bet you take every time, even though you will be wrong nine times out of ten. The point is not recklessness. It is expected value. Once the loop is cheap enough, being wrong stops being the danger. Not swinging becomes the danger.
Slow buys one thing: the comfort of not finding out you were wrong.
Swing.