the cost of building fell through the floor
there’s a sentence i keep hearing lately, some version of “i built this in a weekend,” and it used to mean something and now it barely means anything, because everyone built something this weekend. the wall that used to sit between an idea and a working thing is mostly gone. you can describe a product in a paragraph and watch it assemble itself in front of you, and that is honestly one of the more remarkable things i’ve seen in my working life.
so “can you build it” is answered. solved. yes. you can build almost anything now, and fast, and that isn’t a flex anymore, it’s just tuesday.
which is exactly why being proud of answering it is the wrong instinct.
can was never the bottleneck
here’s the thing nobody wants to sit with: “can i build this” was rarely the real constraint, even before any of this. the constraint was should this exist, will anyone actually use it, will i still care about it in three months, am i the right person to own it, and what happens on the day it breaks with a real person on the other side of the break. none of those got easier. if anything they got harder, because the friction that used to answer some of them for you, quietly, by default, is gone.
and friction was doing you a favor. when building was expensive, the cost itself filtered out most of the things that should never have existed. you didn’t ship the fourth version of something that already existed, because it wasn’t worth three weeks of your life. now it’s worth twenty minutes, so you ship it, and the twenty minutes feels free right up until you remember that building the thing was never the expensive part. owning it is.
everything you ship is a promise
this is the part the weekend-build crowd skips. the demo is free now. the product, i’ve written about this before, was never the demo. but there’s a layer past even that, and it’s the boring, unglamorous fact that anything you put in front of a person is a promise you now have to keep. it will break. someone will use it in a way you never imagined. someone will come to depend on it, and then it will go down at 2am, and that someone is now your problem. capability got cheap. responsibility didn’t move an inch. nobody automated the part where you have to keep caring about the thing after the applause stops.
and that’s the gap, because AI is spectacular at capability and completely silent on judgment. it will build you anything you ask for and it will never once ask whether you should. it has no taste. it doesn’t know that the best engineering decision is usually the feature you refuse to build, the scope you cut, the clever thing you delete because clever was never the point. that judgment is the whole job now, precisely because the building isn’t. the bottleneck moved. it went from “can you make it” to “do you have the taste to know what’s worth making,” and that’s a harder question, and it’s not one you can prompt your way out of.
i’m not innocent here
i should be honest, because this is very easy to write as a lecture and i am not standing outside the thing i’m describing. i have built things i should not have. i’ve shipped features because they were fun to build and not because anyone needed them. i’ve cloned things that already existed, better than my version, told myself it was for learning, and let them rot in a private repo the moment the novelty wore off. what AI changes is the speed, so now you can assemble a graveyard of should-not-have-builts in half the time you used to. moving fast is not the same as moving somewhere. i learned that the slow way, more than once, and i’ll probably have to learn it again.
where i build from
i work in a market where you genuinely cannot afford to build everything, and i’ve come to think of that as a gift more than a limit. when resources are tight, “should you” stops being a philosophy question and becomes arithmetic. you can only carry so many things. you can only maintain so many things. you can only be woken at 2am by so many things before you start asking, before you write a single line, whether this one has earned the right to exist. that constraint used to feel like something holding me back. lately it looks a lot more like the thing keeping me steady while everyone else drowns in stuff they made because they could.
so, should you?
i’m not telling you to stop building. build constantly, build badly, build things nobody asked for, that’s how you learn and i won’t pretend otherwise. this is about one specific delusion: that because the doing got easy, the deciding did too. it didn’t. the deciding got harder, and more important, because there is nothing filtering for you anymore. you are the filter now. that is the job.
you can build anything. congratulations, so can everyone. the only question left that’s actually worth your attention is the one the machine will never think to ask you.
should you?