Why Computers Started Counting From 1970

Open a terminal and type date +%s. The number that comes back is around 1.74 billion: the count of seconds since 00:00:00 UTC on January 1, 1970. That moment is called the Unix epoch, and it is the silent reference point underneath nearly every timestamp on the internet.

The choice was practical, not poetic. Ken Thompson and Dennis Ritchie were building the first version of Unix at Bell Labs in 1969 and 1970. Early versions used a 32-bit integer counting sixtieths of a second, which would overflow in about 2.5 years — embarrassingly soon. They switched to whole seconds and picked a round nearby date that hadn't happened yet when they were writing the code. January 1, 1970 was three months out.

Storing time as a single integer is a beautiful simplification. Comparing two moments is just a subtraction. Adding an hour is adding 3600. Time zones, daylight saving, leap years, and calendar reforms all become formatting concerns layered on top, not data-model concerns underneath.

There's a catch. A signed 32-bit Unix timestamp runs out at 03:14:07 UTC on January 19, 2038. The same kind of overflow that ended the original Bell Labs counter will end the 32-bit one — except this time it's wired into embedded controllers, file formats, and network protocols across the entire planet. Most modern systems have moved to 64-bit timestamps, which last about 292 billion years.

The fix is mostly done. The Y2K of 2038 will be quieter than the original.