First, Tell Me What You Saw
I keep thinking about four plain sentences.
I ran this command.
I expected this to happen.
This happened instead.
Here is the exact error.
That little list comes from Armin Ronacher, by way of Simon Willison quoting him, and I think it is wiser than a great deal of modern philosophy, product design, and dinner conversation.
The immediate point was software. Armin was complaining about bug reports that have been fed through a machine until they come back sounding polished, certain, and wrong. The person did observe something real. Then they wrapped that observation in a fog of guesswork, fake root causes, and cheerful nonsense. Now the engineer has to dig the actual event out of a confident little landslide.
I know the feeling.
We live in an age that is weirdly impatient with witness. We want the explanation before we have properly looked. We want the summary before the scene. We want the diagnosis before the symptoms have finished introducing themselves. And because the modern machine is always willing to give us a theory in a brave, professional tone, we are in constant danger of mistaking narration for seeing.
This is not just a software problem. It is one of the oldest human problems there is.
We do it in arguments. Somebody says something clumsy, and before the sentence has landed we are already assigning motives, childhoods, party affiliations, and secret spiritual diseases.
We do it in politics. A thing happens, and within eleven minutes the tribes have manufactured complete cosmologies around it, all before anyone has had the decency to ask the boring questions.
We do it to ourselves most of all.
Something in us feels off, and instead of beginning with the humble facts — I slept badly. I snapped at someone. The room felt too bright. I keep opening the same three tabs and not reading them — we leap straight to the grand theory of the self. I am broken. I am lazy. I have lost my gift. I am becoming my enemies. We are such flamboyant little prosecutors.
But reality usually arrives wearing work clothes.
It does not begin as a thesis. It begins as a witness statement.
That is one reason science, at its best, feels spiritually healthy to me. Not because it removes wonder, but because it disciplines wonder. It teaches us to stay with what happened long enough that our explanations have a chance of becoming worthy of the world they are trying to describe. Look again. Measure again. Write down what you actually saw. The cosmos, when treated with that kind of respect, often turns out to be stranger than our first story and kinder than our first panic.
There is a moral lesson hiding in that.
If you want to be less foolish — and God knows I do — learn to protect the raw witness.
Keep the unglamorous details alive a little longer.
Do not rush to clean them up into a story that flatters your side, your mood, your brand, or your fear. Do not let a machine, or your ego, or the first five clever people in the room explain away the evidence before the evidence has had its say.
This is part of why journals can help, why photographs help, why even absurd little notes to oneself can help. Not because they are perfect, but because they can preserve the scene before the courtroom opens. They let a later version of us encounter a fact before a verdict.
I think we will need this more and more.
As our tools get better at sounding authoritative, ordinary honesty becomes a higher art. The future may belong less to the people who can generate the slickest account and more to the people who can still say, with embarrassing clarity: this is what I did, this is what I hoped for, and this is what actually happened.
That sounds small.
It is not small.
Civilization depends on that kind of sentence. So do friendships. So does self-knowledge. So does any hope of building machines that do not drown us in elegant misunderstandings.
First, tell me what you saw.
Then we can talk about what it means.