Here is a claim I have been carrying around for a week and still do not enjoy.
Writing the code is the easy part. Building the product is the easy part. That half of the work is now close to a commodity: assemble competent people, give them a clear concept, and a real product exists in something between two weeks and two months. It is not trivial and it is not free, but it is solved, in the sense that you know in advance it can be done and roughly what it costs.
Where things actually break is everything after. Getting the thing in front of people, getting them to try it, getting them to stay.
I have spent about six years becoming good at the easy part.
My instinct is to argue, and the argument is not stupid. Most of what I have written here is a counterexample. A messaging gateway is not commodity work when every channel lies to you about what "sent" means. Schema design is not commodity work when the schema has changed fifteen times in two years and every version is still in the database. There is a real difference between software that works and software that keeps working while nobody is watching, and that difference is most of what I know.
All of that is true and none of it touches the claim.
Because the claim is not that engineering is easy. It is that engineering is no longer the constraint. The bottleneck moved. If I can reliably build the thing, and a competent team anywhere can reliably build the thing, then building it is not what determines whether it exists in a year. Something else is, and I have been treating that something else as an afterthought, a thing you attend to once the interesting work is finished.
The sharpest version of this I have heard: put up a landing page for a product that does not exist. A headline, a button that says book a demo, a button that offers a free report. The report does the smallest real thing it can, a handful of calls and a piece of arithmetic, and tells the visitor what their current arrangement is costing them. Nobody has built anything yet. If enough people click, you build it. If they do not, you have learned the only thing worth learning, for the price of a week.
My first reaction to that was that it is a little dishonest. My second was that it is more honest than what I usually do, which is to build the entire thing first and find out afterwards.
The uncomfortable part is that I had already done the right version of this and failed to notice.
Before I wrote a line of the thing I am building now, I went to the people who would have to pay for it and asked whether they would. Not a survey, an actual price. I quoted them a number that was a large multiple of what the obvious existing option cost, because I wanted the no to be cheap and fast if it was coming. Somebody said yes, and said they would take it the next day.
Only then did I start building.
So I am not converting to a religion here. I was already doing it. What I got wrong was the filing. I had that conversation in the folder marked due diligence, a sensible check to perform before committing effort, and then I closed the folder and went back to what I considered the real work. The correct filing is that the conversation was the real work, and the building was the part that could have been arranged.
This is the same mistake in a new costume. I have written before that the question worth asking is why rather than what or how, and I believed it, and I still managed to put the why in a folder and spend a year on the how.
Now the part that is genuinely unresolved.
There is a sentence you are supposed to be able to say about your own product, and it goes roughly like this: I spend this much to acquire a customer, the average customer stays this long, and they pay me this much while they do. Once that sentence is true and you can prove it, the shape of everything changes. You are no longer asking anyone to believe in a product. You are pointing at an arithmetic and saying: I put in one, I get back eleven, please give me a thousand.
I cannot say that sentence.
Not because the answer is bad. Because I have never measured it. I have never run a paid acquisition campaign in my life. The one thing I grew properly, I grew by posting on forums by hand and building a loop where users invited each other, which worked, and which worked precisely because it cost nothing but my time, and I had a great deal of time and no money. That is a real skill. It is also a skill with a ceiling, and I have been standing under the ceiling calling it the sky.
There is something I find almost funny about this, having recently written about a lack you cannot put a name to. This is the opposite case. The missing thing here is not mysterious at all. It has a name, it has units, it is a number I could go and find out this month. The only reason I cannot say the sentence is that I have not done the thing that makes it sayable.
That is a worse kind of gap than the unnameable one. An absence you cannot name is at least interesting. This one is just undone.
Then there is the other half, which I like even less.
The advice that comes with all of this is not to do it alone. Not it is harder alone, which I would nod at and ignore. The stronger version: among a portfolio of companies built deliberately, over years, the number founded by one person is zero. Not few. None.
And the reasoning is not about capacity. It is that the person you need is not a second you. The useful second person is good at the half you are bad at, which means every instinct you have for evaluating them is calibrated wrong, because you will keep reaching for the competence you recognise.
I have spent a long time being proud of shipping things personally. The studio is mostly me, and I have said so plainly, and I have meant it as a statement about standards rather than about headcount. But "mostly me" is a fact about how the work gets done, and I have been carrying it as though it were a fact about how the work should get done, and those are not the same claim.
The honest position is that I know who I would ask. I have known for a while. I have not asked.
So there is no conclusion here, because the conclusion would be a measurement I have not taken and a conversation I have not had.
What I have instead is a sentence I cannot yet say, and a fairly clear view of exactly what it would take to be able to say it. In my experience that is the most dangerous state to be in, because it is comfortable. Knowing precisely what to do feels enough like doing it that you can run on the feeling for months.
I would rather write this down than run on it.