Almost the right question.
I watched TOTO play in Sweden. At one point they asked whether anyone in the audience fancied a go, and a man climbed up and almost played the Rosanna shuffle. Almost turns out to be the interesting part of most things.
He was not bad. He had clearly practised. He got the shape of it, and for a few bars you could hear the song trying to happen. The band grinned at each other, the way people do when someone reaches for the hard thing in front of a few thousand people. Then they let him keep playing, and held the time underneath him so the song stayed up.
What is actually difficult
The pattern is not hard to describe. A shuffle, played in half time, with quiet strokes filling the space between the loud ones. You can read it off a page in a minute and play the notes by the weekend.
What you cannot get from the page is the quiet strokes. They are barely audible on purpose, and they are the entire reason the loud ones land where they land. Nor can you get the time: the feel that sits fractionally against the click and makes the thing breathe. Play every note correctly and slightly wrong in time and you have played the notes and not the song. Everyone in the room hears the difference and almost nobody can name it.
That is the shape of the gap. Not the part you are looking at. The part you did not know was doing the work.
The counter
These days I am the man behind the counter of the corner shop. Head of engineering, head of data, head of the whole store, one room. People come in all day and ask me where things are, and the questions are almost right.
That is not a complaint. It is the most useful information I get. The question someone brings is an exact map of what they can already see, and therefore of where the edge of their experience runs. From four recent weeks:
"Which model should each of my scheduled scripts use?" The real question: why is a model running your scheduler at all? Most of those jobs were not thinking about anything. He was optimising the fuel of an engine that did not need to be there.
"How do I keep the API key out of my app?" The real question: why does one key open your entire warehouse? Hiding it is a good instinct aimed at the wrong layer. The exposure is not where the key sits, it is what the key permits.
"Is a file in a sync folder good enough for my ledger?" The real question: what guarantees a payment is written exactly once? Storage was the visible part. Uniqueness under two writers was the part doing the work.
"How do I make this report faster?" The real question: which of its eleven layers can you delete? Speed is a symptom. Each layer was reasonable on the day it was added, and that is exactly why nobody can point at the guilty one.
In every case the person was competent, the thing they built worked, and the question was aimed one layer above the problem. That is not ignorance. It is the horizon of what they have personally paid for.
Experience is not knowing more things. It is having already paid for these specific mistakes, and recognising the shape of them early.
Nobody invents from nothing
The man who wrote that drum pattern was open about where it came from: he took a shuffle he had learned from one drummer and a shuffle he had heard from another and made a third thing that outlived him. That is what experience actually is. Not talent, and not gatekeeping. Inheritance, plus repetition, plus a long memory of what did not work.
I asked every one of the four questions above, in some form, on slower machines with nobody to ask. My mistakes were smaller only because the systems were smaller and the feedback took months. Nothing about being early makes you clever. It makes you someone who has already been wrong in a specific way, which is a cheaper thing to buy than to repeat.
Keep playing behind them
So: the more the merrier, sincerely. The barrier came down, a lot of useful things now exist that nobody would ever have commissioned, and the people building them are having the time of their lives. I am not interested in telling anyone to get down off the kit.
But somebody in the room has to know which question you meant. Not to take the sticks away. To say: the quiet notes are doing more than you think, you are rushing the third bar, and the thing you built which works beautifully today will stop working in November for a reason I can already name.
That is the job now, and it is a better job than it used to be. It is also the one part that cannot be generated, because it is made entirely of having been wrong before.