We used to settle for thumbnails.

Something quiet just changed.

For thousands of years, institutions ran on human attention, and human attention has limits. A teacher with thirty students. A doctor with forty patients a week. A college advisor with three hundred names on her caseload. To hold each of those people at full resolution would have taken billions of hours. So every institution that served you compressed you into a thumbnail. Then it learned to act from the thumbnail.

You can think about it the way photography went. We didn’t always have high-resolution images of everything. We had paintings, and only royalty could afford the full likeness. We had early photographs, but they were precious and rare — one shot per occasion, no retakes. Then film got cheap, then digital, and then storage and transmission got so cheap that we stopped thinking about it at all. Today your phone holds tens of thousands of full-resolution images and the question is just which ones to look at.

The same thing just happened to people inside institutions. The price of the full picture has collapsed.

The choice.

So every institution has two ways to spend the discount.

It can use the discount to sort you faster. More candidates screened, more students flagged, more patients triaged. Same thumbnails: just more of them, generated more quickly. That’s the obvious move. Efficiency has a clear ROI. And the metrics most institutions already track (volume, throughput, latency) reward more of what they’re already doing.

Or it can use the discount to hold you at higher resolution. Same number of students, same number of patients, same number of candidates. But what the advisor opens on her screen is no longer a thumbnail. It’s the picture.

The CEO who says she needs the thumbnail because attention is scarce isn’t wrong about the constraint. Human attention is still scarce. The picture itself isn’t an asset unless someone designed it for the moment of decision. A full-resolution view dumped into twenty minutes without curation is overload, not insight. The discount has to buy two things: the picture, and an interface that lets the advisor use it in twenty minutes without drowning. The question is not thumbnail or full picture? It is what does the full picture need to look like for the advisor to use it in twenty minutes? That’s a design problem. It’s the second half of what the discount buys.

Most institutions are reaching for the first option. It is what the dashboards already measure, what procurement already buys and what the consultants already pitch.

The second option requires a different question.

The question.

The question is one a measurement scientist would recognize and one an advisor can carry into any room: does this system actually see the person, or is it still working from a thumbnail?

This is a measurement standard. The discipline is called validity, and it has decades of track record from the world of high-stakes testing. Its central claim is stricter than it sounds: a score is never valid on its own. What gets validated is a particular reading of that score, put to a particular use, judged partly by what happens to the people it lands on (Messick, 1995). A number that tracks length, fluency, demographics, or the candidate tested just before you fails that standard twice. It isn’t measuring what it says it is, and the decision made with it isn’t the decision anyone signed up for. The standard isn’t new. It’s been ready for a while.

What’s new is the price of asking it. Validity has always been applied to scores, and scores are thumbnails. When carrying the full picture was impossible, the picture itself could not be validated. Now it can. And now it should be.

The stakes.

For most of history, the thumbnail was the kindest thing an institution could do for you. Institutions did not withhold the full picture out of ill will. They withheld it because carrying it was impossible.

The question for 2026 is whether institutions use the savings to generate more thumbnails faster, or to finally stop settling for them.

We used to settle for thumbnails.

We don’t have to anymore.

Written July 2026 for the Analytic Bytes Library. The argument here is compact by design; the longer versions live in companion pieces: What is this system actually measuring? (the validity question applied to an institutional system), The numbers don’t agree because the words don’t (why two systems that both claim to describe the same person disagree), and The Decision System (what the full picture has to look like to inform the twenty minutes without overwhelming them).