When we think of a cache, we think of a way of storing information "closer" to the place it's needed. Most general-purpose CPUs, for example, have an on-board cache which is used to avoid accessing main memory - after all, the memory that's on the same die as the CPU is going to be quicker to access than the RAM chips. Filesystems use caches of RAM to make disk accesses appear to be quicker, as RAM chips are much faster than moving a mechanical arm across a spinning disks. If we're lucky, and if we've got a good caching algorithm, we can get an impressive speed boost by keeping the right bits in RAM. Likewise, CPUs get a speed boost by keeping the right instructions and data on chip.

Caching is not limited to CPUs and filesystems, of course. Most browsers maintain a cache of pages, of images, of css files, of javascript, and of any other bit of information that is useful for displaying web pages. By using a local on-disk cache (some of which is going to be in RAM anyway, thanks to the filesystem), browsing appears to be much quicker than it would if the browser had to re-load every single image from a distant web site. The browser does check to see if any files need to be retrieved again (see http's 304 Not Modified response code), so there is some over-the-wire activity, but that's about it.

All of this is a long-winded way of saying I was amused by the second bullet item here (from Apple Insider):

Either the cache is poorly implemented, or the users reporting this information are confused.