Moore's law
Moore's law is the observation that the number of transistors you can pack onto a chip of a given cost roughly doubles every two years. Picture a city block that holds twice as many houses each time you revisit it, while the rent stays flat — that relentless doubling is why the phone in your pocket dwarfs the room-sized computers of the 1960s. It is a trend about economics and manufacturing, not a law of nature like gravity.
Gordon Moore, a co-founder of Intel, sketched the pattern in a 1965 magazine article from just a handful of data points, and the industry then spent decades treating it as a roadmap to hit. Each doubling came from shrinking the features, so transistors got cheaper, faster, and lower-power all at once. A 1971 chip held about 2,300 transistors; a modern processor holds tens of billions. The pace is now slowing as features approach a few atoms wide and tunnelling, leakage, and soaring fab costs bite.
Moore's original 1965 paper actually predicted doubling every year; he revised it to every two years in 1975, and the popular "18 months" version came from Intel executive David House, who folded in speed gains as well. Don't confuse Moore's law (transistor count) with Dennard scaling (power density), which broke down around 2005 — that breakdown, not Moore's law itself, is why clock speeds stalled and the industry pivoted to multi-core.