What deserves to exist when software is cheap to produce?
AI has shortened the distance between an idea and working software. I am still learning what should happen before we start running.
A thought, held still long enough to examine.Software used to make an idea pay a kind of entrance fee. Before anything useful appeared on a screen, someone had to decide the problem was worth days or weeks of setup, implementation, and repair. That friction killed good ideas along with bad ones, but it also forced a question early: is this worth building?
AI has moved that question. I can now test an interface, compare approaches, or turn a small irritation into a working tool much faster than I could a few years ago. I like that. I also notice how easily motion can begin to feel like progress.
The bottleneck moved
Implementation still matters. Generated code can be wrong, insecure, difficult to maintain, or simply unrelated to what a person needed. But the scarce part is moving upstream. The harder work is increasingly deciding whose problem is being solved, what the product should make clear, and which possibilities should be left out.
A fast prototype can answer whether an idea can exist. It cannot answer whether the idea deserves someone’s attention, trust, money, or time. Those are judgment questions, and speed does not make them disappear. It makes them arrive sooner.
From the marginAI has made it much easier to build software. It has not made it easier to decide what deserves to exist.
A real person is a useful constraint
The independent projects I care about most begin with someone specific. The RuneScape alchemy tool began with a comparison I wanted beside the game. The quieter games site began because my friends wanted to spend time together without returning to another intense multiplayer night.
Neither idea needed a total addressable market slide before it could be useful. The audience was small, but I could see the moment the product was meant to improve. That creates a useful constraint: if I cannot explain what changes for that person, more features will not rescue the idea.
The question I want to keep
The wrong conclusion would be that small tools are meaningful and ambitious products are suspicious. I want to build ambitious things. I just do not want ambition to become an excuse for producing software whose only clear achievement is that it was fast to produce.
So I am trying to keep one question near the beginning: if this works exactly as imagined, what becomes meaningfully better for another person? I will not always know the answer before building. Sometimes the prototype is how I find it. But if the answer never becomes clearer, stopping may be the most honest product decision available.