The unglamorous middle of the day
Ask where artificial intelligence has actually earned its keep and the honest answer sits in the middle of an ordinary working day. Sorting a pile of things into an order. Naming files so they can be found again. Pulling the three relevant lines out of a long document that nobody was ever going to read all of.
None of that makes a good demonstration. It is the sort of work that was previously done badly or late, by someone who had meant to get to it, and the gain is measured in recovered afternoons rather than in anything that can be shown on a stage.
It is also where mistakes are cheap. If a file is named oddly it gets renamed. That low cost of being wrong is exactly what makes the middle of the day a sensible place to start, and it is the opposite of the places that get the attention.
Why a first draft is the honest use
The second quiet use is the blank page. A draft produced in a few seconds is rarely good, and that turns out not to matter very much. Its job is to be wrong in a specific, visible way, because reacting to something wrong is far easier for most people than inventing something from nothing.
The draft is not the finished thing and should never be mistaken for it. Anyone who has edited knows the difference between a page that needs work and a page that does not exist, and only one of those two problems is solvable on a Tuesday afternoon.
What this asks for is a particular habit: treating the output as a proposal rather than an answer. That habit is unfashionable because it is slow, and it is the single thing that separates useful work with these tools from work that merely looks finished.
The test is whether anyone notices
A tool that is genuinely working becomes invisible. Nobody praises the thing that quietly removed forty minutes from a Thursday, because the forty minutes were never visible in the first place. The test of usefulness is whether the absence is noticed when it stops, not whether the presence is admired while it runs.
That is a very old pattern. Most of the technologies that reshaped ordinary work did it without ever becoming a topic of conversation, and the ones that stayed a topic of conversation often turned out to be doing less than they claimed.
None of this is a forecast, and none of it is a recommendation about what anyone should adopt. It is simply where the useful part seems to be at the moment: low down, close to the ground, and considerably less interesting than the version in the headlines.
Frequently asked questions
What counts as an everyday use?
Small repetitive work in the middle of an ordinary day: putting things in order, naming them so they can be found, and reducing something long to the few lines that actually matter.
Why is a rough first draft useful if it is rough?
Because reacting to something wrong is much easier than inventing something from nothing. The draft earns its place by being specific and visible, not by being right.
How would anyone tell whether it is working?
By whether its absence is noticed. A tool that is genuinely helping becomes invisible, and the time it saves was never visible to begin with.