BUILDER NOTES
AI can take you from 0 to 80. Builders still matter from 80 to 98.
AI has made the first version of a product dramatically easier to create. It has also made one thing clearer: the density of the final product still comes from the builder.
A note from watching AI startups closely, and building several products ourselves.
Fast first version
The strongest AI zone. Ideas become screens quickly.
Structural understanding
You have to understand what the model produced.
Product judgment
Conflicting details, use cases, and constraints need to be resolved.
Builder edge
The tool fades. Taste, architecture, and patience show up.
Watching AI startups closely, and building several products ourselves, I keep coming back to the same observation.
AI is extremely useful at creating a fast, plausible first version. It is excellent at taking an idea from 0 to 80. A landing page appears. A prototype works. A demo becomes clickable. The distance from thought to artifact has collapsed.
That speed is intoxicating. Builders start producing A, B, C, and D at the same time. I have done the same. But the problem starts after that. Most things stop at 80.
After 80, building becomes understanding
Moving from 80 to 90 is much harder than going from 0 to 80. At this stage, using AI is no longer enough. You have to understand what AI produced. You have to see the structure, identify the weak parts, correct what is wrong, and decide what the product is actually trying to become.
The next stage, 90 to 95, is harder again. What looks like a small gap from the outside is usually a full pass through product logic, flow, performance, data, user experience, and operational reality.
Most AI-built products cluster between 0 and 80, sometimes near 90. In that range, products start to look similar, compete with each other, and disappear into market noise.
Only a small number reach 95. That is when a real relative advantage begins to appear. The question is no longer who made something fastest. It is who understood it deeply enough to finish it.
After 95, the person becomes more visible than the tool
Moving from 95 to 96, and then 96 to 97, is a different kind of problem. Effort alone is not enough. The builder's ability to use the tool, the quality of the idea, the structure of their thinking, and their design judgment all become decisive.
The final result does not come from the tool. It comes from the builder's mind. Michelangelo did not create sculpture by randomly moving his hands until something emerged. The finished form already existed in his head.
AI makes the hand faster. It does not decide what should remain, what should be removed, how the structure should hold, or how far the builder is willing to take responsibility.
Working note
Speed of production is not the same as product density.
It is genuinely enjoyable to watch AI turn ideas into artifacts. But with some distance, many of those artifacts are closer to byproducts of production than products that should exist in a market. There is a lot of evidence that something was made. There is much less evidence that it was understood.
A very small number of people have a clear picture. They know what should be built, how it should be built, and what has to be pushed through when the easy part is over. They create the one or two percent difference that proves why they are in the loop.
At 97, whether AI was used becomes the wrong question
Something interesting happens once a product reaches 95 or beyond. The question, was this made with AI, starts to matter less.
At 80, the question is valid. Without AI, many of those things would not have been made. But at 97 or 98, the question becomes less useful. Even if AI was used, not everyone can reach that level.
So the question I keep coming back to is not how much can we create with AI. It is what can we understand deeply enough to own. How far can we take responsibility. Can we keep going after 80, and create the small differences that matter after 95.
AI changed the starting line. The density of the final product still comes from people.