You do not get good by getting it right. You get good by getting the reps.
The last five percent is a trap, consistency beats intensity, and a failure is a data point rather than a verdict. A companion to the personal-development piece, about the one relationship most founders need to change: the one with being wrong.
You do not get good at something by getting it right. You get good by getting enough reps that getting it right becomes cheap. This sounds obvious and then almost every founder violates it, either by polishing one thing to death or by abandoning it before it ever worked. Both are the same mistake wearing different clothes: avoiding the boring middle where the reps happen.
The last five percent is a trap
Getting to eighty or ninety-five percent of good is fast. The final five percent takes more effort than everything before it combined. Anyone who has shipped software knows the shape of that curve. The same curve governs a new hire, a new skill, a new process. At ninety percent you can already start capturing the benefit. Chasing the last five now burns the exact time you needed for the next rep, and the next rep is worth far more than the polish.
What the reps are actually doing
Consistency beats intensity
A heroic week you cannot repeat is worth less than four good days out of seven, sustained. Intensity feels like progress because it is exhausting. Consistency is progress, precisely because you can do it again tomorrow. If you are recovering from a stretch that broke you, cap the workload so the recovery holds, the way you would after any real setback. Do less, sustainably, and let the compounding do the rest.
A failure is a data point
A thing that broke is information you did not have yesterday. That is not a consolation, it is the actual mechanism. The iterations are the struggle and the struggle is what builds the skill. Enough reps of getting it wrong and you start getting it right first time, because you have seen every way it fails. So when something breaks, the useful question is not who to blame, it is what this just taught you and what protocol stops it recurring.
Some work is a learning gig, not an earning gig
A project that pays badly or moves slowly but teaches you a capability is not a bad deal, it is tuition. You are paid in a skill you did not have, and next time you charge for it. Read a hard, underpriced piece of work that way and it stops being a drain. It is the cheapest way to buy a new thing you can sell.
The other failure mode: never finishing a rep
The opposite of over-polishing is shiny-object syndrome: dropping the current thing for a new idea before the current thing ever came to fruition. It feels like progress because a new idea is exciting and an unfinished one is boring. But you do not get the reps if you never let one play out. Pick something meaningful, stay less rigid about how, and stay committed to seeing it through. You are not going to run out of chances. You are going to run out of finished things.
Ship at good enough, keep the cadence you can actually hold, treat every break as a data point, and finish the rep. That is not a productivity hack. It is a different relationship with being wrong, and it is one of the most personal changes a founder makes.
Stuck polishing the wrong thing, or never finishing?
A lot of what I do is help a founder pick the rep that matters and hold the cadence to finish it. The strategy call is free and it is a real diagnosis, not a pitch. If the timing is off, the community costs nothing.