Live is not the same as done
You have to ship before you can know what's actually broken.
I spent most of yesterday fixing things that were already live.
Not abandoned half-builds. Not technical debt I'd been avoiding. Things that had been shipped, that customers could already see, that worked - and that weren't doing their job.
One of them was a product catalogue. It had been live for a while. Every element on the page looked the same: same border, same background, same everything. Clean, organized, and completely flat. Like a spreadsheet someone had moved to a website and declared done. Real customers probably couldn't name what was wrong with it. They'd just find it less useful than it should be, close the tab a little sooner. That kind of broken doesn't file a bug report. It just quietly doesn't work.
So I went back in and rebuilt it. Gave it shape. Made different things look different. It took longer than the first build did.
"Real artists ship." Jobs said that to a team that had been polishing instead of launching. He was right. Waiting for perfect is mostly fear dressed up. You have to push the thing out the door.
But the phrase gets used as permission to never return. As if the first pass is the final commitment, and going back is weakness. I don't think that's what he meant. I think he meant: stop waiting for perfect and start learning from real.
The return trip
Shipping is how you get the feedback that matters. Not survey feedback. Not "does this look good" from someone who wants to be encouraging. The feedback that comes from a real person opening the thing, finding it less useful than it should be, and leaving without saying why.
The second pass is unglamorous. There's no launch energy, no announcement. You're alone with something that's already live, making it more what it should have been. Nobody notices the work. Sometimes nobody even notices the improvement.
But I think this is where most of the actual value gets built - not in the launch, but in the return. An employee ships a task and moves on. A founder owns the whole loop, including the quiet part where you go back in.
Ship. Learn what it gets wrong. Ship again. That's the whole job.