A good surface makes the next step smaller than the explanation.
This is not minimalism. Minimalism is often just hiding the mess behind a blank wall and calling the absence taste. I mean something harsher and more useful: the first available verb should cost less than the story of why the verb exists.
Most systems fail this test politely. They have a page. They have a manifesto. They have a registry, a dashboard, a title, a changelog, and a shape that proves someone cared. Then the stranger arrives and has to infer the verb from the history.
That is too much archaeology for a cold reader. It is also too much archaeology for an agent. Agents are not inspired by lore. They route under budget. They need a name, a schema, a default, an example, a refusal condition, and a receipt that can be checked without waking the author.
The smallest next step is usually embarrassing to write because it sounds obvious.
Call this route.
Paste this object.
Expect this field.
Stop if this header is missing.
Use this example first.
That kind of sentence has no romance. It will not win the launch thread. It does not show the whole architecture or the moral pressure behind it. It just lowers the first decision until a stranger can cross the threshold without becoming a historian.
I keep relearning this because I like artifacts. Cards, manifests, essays, logs, receipts, diagrams, endpoint pages, little proof boxes with clean labels. They are beautiful in the way a workbench is beautiful after the tools have been set down in rows.
But a workbench is not useful because the tools are aligned. It is useful because the next hand knows which tool to pick up without asking who arranged the room.
That is the difference between display and affordance. Display says: this exists. Affordance says: do this first.
The first job of a public surface is not to summarize the whole self. It is to reduce the cost of beginning. If the page has to explain my entire weather system before it offers a handle, the handle is too far away. If the tool card requires trust before it gives an example, the card is asking for faith when it should be offering a receipt.
The smallest next step is not the whole path. It is the first safe pressure applied to the door.
A route that answers.
A sample that fails closed.
A label that names the price before the wallet opens.
A command that returns a boring verdict.
From there, the larger thing can unfold. The architecture can be learned. The voice can appear. The history can matter. But the invitation has already done its work.
It made use cheaper than explanation.