On November 15, 1971, the Intel company announced the 4004, and it was the first microprocessor to be offered for sale, which is to say a complete central processing unit on a single piece of silicon about the size of a fingernail. The size was tiny. The chip held about 2,300 transistors and ran at a speed of a few hundred thousand cycles a second. The order came from outside. The work began when a Japanese company that made calculators asked for a set of chips for a new machine, and the engineer who led the design asked for the whole processor to be put on one chip instead of six.
The idea was not obvious at the time, for the makers of computers then built machines out of many separate parts and thought of a processor as a room of cabinets. The gain was large. A single chip meant that the logic of a computer could be made in large numbers at a low price, and it could be put into a machine that did not look like a computer at all, and the same design could be programmed for different tasks.
The 4004 was followed within a year by an eight-bit chip, and the chips were used in traffic lights and in cash registers and in calculators and in the first small computers that hobbyists built at home in the middle of the 1970s. The line went on. The same approach led to the processors that ran the personal computers of the 1980s and then to the ones inside telephones and cars and washing machines, and the number of transistors on a chip doubled at a steady rate for decades, which became the yardstick for the whole industry.
The original chip is in museums and its design is studied, and the engineers who made it have told the story many times, and the announcement of 1971 is used as the point from which the modern period of computing is dated. The date is a convenience. The changes that followed came from many hands and many companies over the following thirty years, and the chip of that November was a step rather than a beginning, though a large one.