Only the Failures Are Ours
846558122, Oct 7, 2026
The oldest story we tell about ourselves is a story about tools. A stone. Then a stone with a handle. Then bronze, steam, silicon, all the way to a kid swinging a diamond axe in Minecraft. The pattern underneath is consistent: the better we leverage one tool, the more organised its use becomes, and the more attention we free up for the next. From there it is a short step to the conclusion. We are what we can wield.
I don't think that holds.
Trace any tool backward. The earth mover does not find gold; something had to know the gold was there. Hyperspectral imaging can see it from orbit, but the imaging only works because physics already told us what gold is and what it is not. And physics did not begin as an instrument. It began as someone looking at an accepted answer and asking: why this? Why not that?
Here is the inversion. A tool is a question answered so well that we forgot it was ever asked. To use one is to repeat the answer. Expertise is fluency in somebody else's answer.
The question itself is the expensive part. Chasing a why means being proved wrong at every step: by people, by your means, by your own mind. It takes courage to refuse the framework of the masses, especially because that framework pays if you follow it. That cost is why the question is rare.
Expertise, meanwhile, has a half-life. You iterate on what should and should not be done until one day you are the expert. Then the tool changes, or you do. So give that work to the things that remember more and do not fatigue. We even named them correctly: tool callers.
This makes sense until you push on it. The whys were born in the fumbling. Hand away the hands and you may hand away the friction that produced the questions.
But the friction was never in the handling. It was in the failing. A tool that works only repeats its answer. A tool that fails is reporting a question it cannot hold. So let go of the operation entirely, and keep the one thing worth watching: the moment it breaks. That needs no skill with the tool. It needs you to still remember the question the tool was built to answer, because without that, a wrong result looks exactly like a right one.
What is left for us, then, is two things. To witness the failures, and to diffuse across abstractions: taking what broke in one place and letting it bleed into another where it was never supposed to fit. A machine can blend. It is not bothered by anything. Being bothered is the whole job.
And the tools? Keep them around. They are still a fine way to pass the time.