Every student has made the argument, usually around age eleven: I got the right answer, why do I have to write out all the steps?

The answers they get back are unsatisfying, and mostly true. Partial credit, so a single arithmetic slip doesn't cost the whole problem. Proof you didn't guess. Practice at the notation you'll need later. All real, all sounding like bookkeeping — which is why the argument keeps getting made.

The better answer is one teachers rarely have time to give: writing the steps isn't a record of the thinking. It is the thinking.

The steps are where the reasoning gets forced

You can hold a vague sense of a solution in your head — a feeling that it probably works out, a fuzzy picture of where it's heading. That feeling is durable precisely because it's vague. Nothing in it has to be checked.

Writing it down destroys that comfort, usefully. A line of work has to commit: this specific quantity, this specific operation, following from that specific previous line. You can't write "and then it simplifies" — you have to say to what. Vagueness that survives happily in your head dies on contact with a page, and where it dies is exactly where you didn't understand the thing.

That's the mechanism. Not neatness, not documentation. Externalizing a chain of reasoning makes it inspectable, first to you and then to anyone else. Students who write out their steps discover their own gaps mid-problem, which is the cheapest possible moment to discover them.

And then the answer became free

For most of the history of school, the final answer was decent evidence. Not perfect — you could copy a friend's, or the back of the book — but producing a correct answer to a novel problem usually meant something happened in your head.

That inference is gone. A correct final answer to almost any homework problem is now available instantly, to anyone, for free. Not as cheating that requires effort or nerve, but as the path of least resistance, available on the phone already in the student's hand.

So the artifact schools have collected and graded for a century has quietly stopped carrying information. A page of correct answers tells you nothing about the person who submitted it. This isn't a discipline problem to be solved with better proctoring; it's an evidence problem. We were measuring the wrong thing and got away with it for a long time because the wrong thing correlated with the right one. It doesn't anymore.

The answer was never the evidence. It was just the easiest thing to check.

What's left is what always mattered

The reasoning is what survives. Not because it's impossible to generate a plausible-looking derivation with a machine — it isn't — but because the act of producing a chain of steps yourself is the act of learning, and a real page of real work carries the marks of it. The false start. The branch crossed out. The place where the handwriting slows down and gets careful. The step someone skipped because they thought it was obvious, revealing whether it actually was.

None of that appears in a final answer, and none of it can be inferred from one. It's the difference between an artifact that shows a mind working and an artifact that shows a result was obtained.

Which means the old instruction was right the whole time, for a reason that's only now obvious. Show your work. It was never about partial credit. It was about the fact that the work is the only part that was ever yours.

That's the assignment Lune Synth™ is built around: do it by hand, scan it in, and get graded on the reasoning you actually produced. Everything else is checking answers, and answers stopped being worth checking.

If that's the kind of feedback you want, join the beta waitlist on the Lune Synth home page, or write to griffin@lunesynth.com.