tl;dr: Progressive JPEGs store images as multiple "scans" of increasing detail, and by concatenating multiple images while stripping certain markers, you can create a single JPEG that displays different frames as it loads over a slow network. Since most decoders bail out after ~9 scans, the trick is to use minimal DC-only scans (yielding 1/16 resolution frames), which allows packing ~90 frames—enough for a crude video—into one standards-compliant JPEG. Playback timing depends entirely on network speed, so it has no real practical use beyond novelty demos like HTML-only video via `<dialog>` tags.