Look at a photo of a busy street. Now look at the same street a moment later, with one thing different: a lamp moved, a shop sign recoloured, a car gone. You would spot it at once, right? Most people don't. If a blink of blank screen sits between the two pictures, even large changes can stay hidden for many seconds.
This is called change blindness, and below is a way to feel it. The scene flickers between two versions with a short grey blank in between. One object differs. Tap it.
The picture alternates about twice a second. If flicker bothers you, skip the interactive: the text below covers the same ground.
Now switch the blank flash off and press New scene. The same kind of change, with no blank in between, jumps out almost instantly. Nothing about the change got bigger. What you removed is the blank, and the blank was doing all the hiding.
Normally a change in the world comes with a tiny bit of motion. Something moves, flickers or appears, and that motion signal drags your attention straight to it. You never have to compare anything, because the change announces itself.
A blank screen floods the whole image with motion at once. Every object produces a signal, so the one real change is lost in the noise. Now you have no alarm to follow. To find the change you have to do something much harder: remember how each object looked, then look at each one again and compare.
The experiment that made this famous used exactly this trick. In 1997 Ronald Rensink, Kevin O'Regan and James Clark showed people photographs that alternated with a modified copy, separated by a brief blank. Even large changes to the scene often took many cycles to find.
It feels as though your head holds a detailed picture of the room. It does not. Only a small patch of the retina, the fovea, sees sharply: about the width of your thumbnail at arm's length. Everything else is blurry, and you build the feeling of a full picture by moving that patch around.
What you carry from one glance to the next is thin. Visual working memory holds roughly three or four objects. When the blank wipes away the motion cue, you can only notice a change in the objects you happened to attend to and store. For everything else, the old and new versions look equally fine, because nothing was stored to compare against.
That is why the change feels obvious once you see it. After you have found it, you attend to it, and the difference becomes impossible to miss. The feeling that it was always there is exactly backwards: the picture did not change in your head, your attention did.
You do not need a flickering screen. In 1998 Daniel Simons and Daniel Levin had an experimenter stop pedestrians on a university campus to ask for directions. While they talked, two workers carrying a door walked between them, and during that moment the experimenter was swapped for a different person. Fewer than half of the people noticed that the person they were talking to had changed.
The door plays the same role as the blank: it breaks the continuous view so there is no motion to follow. The people were not careless. They were doing what vision always does, attending to the task (giving directions) and trusting the world to stay put.
This is the practical lesson. Seeing is not recording. If something matters, such as a proofreading error, a road sign or a detail in a photograph you need to compare, look at it on purpose and compare it directly, rather than trusting that you would have noticed.
Run a few rounds with the blank on, then a few with it off, and compare your two times above. For many people the gap is huge. Try a trick too: instead of scanning randomly, pick one region and keep your attention on it for a few cycles. The change often shows itself the moment it is inside the spotlight.