You're looking for fireflies in a dark forest. Instead of photographing the whole forest every second and then poring over the images, you carry a small device that spots the flashes on its own and cuts out only a circle around each one. The rest, the blackness, isn't saved. That's exactly how the new sensor for scientific cameras works.
Typically, high-speed cameras shoot a million frames per second, and each frame is a huge amount of data. Sending it all over wires is impossible: the channel gets swamped. But in most tasks, only rare bright points matter: X-ray flashes when studying crystals, particle tracks in an accelerator, glowing proteins under a microscope. Everything else is empty background. The new chip does something simple: it finds those points itself and sends out only small squares around them — about the size of a postage stamp, not the whole frame.
Tests on real hardware showed that the chip finds all the same peaks as a regular computer, but does it hundreds of times faster and right inside the camera. On an old 130-nanometer process, it can handle hundreds of thousands of frames per second, and on a more modern one, over a million. This opens the door to experiments where data used to be too much to save in time.
The fastest chip tested processes more than a million frames per second — in one second, it can look at as many images as you'd see in a regular movie over five hours.
In The Matrix, characters see a stream of symbols instead of reality — the new sensor also discards everything extraneous and keeps only the meaningful flashes, but it does so without a human.