AI made it cheaper for me to build. It also made validation more important.
I’ve been thinking about something that changed for me after building with AI.
The first instinct when AI makes development faster is to build more.
You can go from an idea to a working prototype surprisingly quickly, so the temptation becomes:
“Since I can build it quickly, why not just build it and see what happens?”
But I think that reverses the order of the problem.
When building was expensive, founders were forced to think carefully before committing engineering time.
Now the cost of building a prototype can be much lower.
That means the scarce resource increasingly becomes knowing what is worth building in the first place.
I’ve started thinking about validation as a sequence of questions rather than a single “is this a good idea?” test:
Who actually experiences the problem?
How are they solving it today?
What makes the current solution frustrating or expensive?
What assumption would have to be true for my product to work?
What is the cheapest experiment that could prove or disprove that assumption?
The interesting part is that AI can help with several of these steps — research, analysis, prototyping, competitive exploration — but it can also make it dangerously easy to produce convincing answers without actually validating the underlying assumption.
That distinction has become one of the things I think about most while building my own AI product.
Has AI changed the way you decide what is worth building, or has it mainly changed how quickly you can build it?
Replies