The Apollo Guidance Computer that landed humans on the Moon had less processing power than a modern microwave, and the engineers programmed it with rope memory that was hand-woven by women who were called "Little Old Ladies" in the official documentation — and the entire system worked because they were never wrong

The standard cultural framing of the Apollo Guidance Computer tends to emphasize one half of the story. The half it emphasizes is the comparison to modern computing. The Apollo Guidance Computer had, by every available measure, almost no processing power by current standards. It had 4 kilobytes of RAM. It had 72 kilobytes of read-only memory. It ran at 0.043 megahertz, which is several orders of magnitude slower than any consumer electronic device currently in production. The standard comparisons note that a contemporary smartphone has roughly 100,000 times the processing power of the AGC, and more than a million times its RAM. A modern microwave, in any honest accounting, has more computational capability than the system that landed humans on the moon.

This is the half of the story the wider culture has absorbed. The half it has not absorbed is more interesting, on close examination, than the half it has. The half it has not absorbed is how the software that the AGC ran was actually loaded into the computer. The loading was not, on the available evidence, accomplished by any of the methods the contemporary register associates with computer programming. The loading was accomplished by women, sitting at long tables in a Raytheon plant in suburban Boston, physically weaving the software into the hardware with copper wire, by hand, one bit at a time.

What rope memory actually was

The technology was called core rope memory. The technology was a form of read-only memory, in which the bits of a program were encoded by the physical path of a wire through a matrix of small ferrite cores. If the wire passed through a particular core, the bit at that position was a one. If the wire passed around the core, the bit was a zero. The entire program was, in this way, physically embodied in the wiring of the memory module. The program could not be edited once the module was wired. The program could not be patched in flight. The program was, in some real way, the hardware.

This sounds, in 2026, like a piece of impossibly primitive engineering. The sounding-primitive misses what the technology was, in fact, doing. Rope memory was, on close examination, one of the more remarkable solutions to a particular set of constraints. The constraint was that the AGC had to fit into a single cubic foot of spacecraft volume, weigh no more than seventy pounds, operate reliably in the radiation environment of space, and store enough software to run the entire guidance and navigation system for a manned lunar mission. The semiconductor memory technologies that would eventually replace rope memory did not yet exist at the storage densities the AGC required. Rope memory was, in some real way, what could be built. Rope memory could store about 72 kilobytes of program in a single cubic-foot module, which was, by the standards of 1965, an extraordinary density.

The density came at a cost. The cost was that the memory had to be wired by hand, by people who could thread copper wire through a precise sequence of thousands of cores without ever making a mistake.

Who actually did the work

The work was done at the Raytheon plant in Waltham, Massachusetts. The workers were, in most cases, women, many of whom had been recruited from the local textile industry and from the Waltham Watch Company, which had previously employed many of them in precision manufacturing roles. The history of who did this work is, on examination, considerably more substantial than the cultural register has tended to acknowledge.

The women sat across from each other at long tables. Each woman would, in turn, pass a hollow needle containing a length of copper wire through the matrix of ferrite cores in front of her, in a sequence that had been specified by the programmers at MIT’s Instrumentation Lab. The needle would pass through some cores and around others, encoding the binary pattern of the program one bit at a time. The wiring of a single rope module took, by some accounts, approximately eight weeks of work. A complete AGC contained multiple modules. The total wiring effort for a single mission’s flight software was a matter of months of skilled labor by trained workers.

The work was, by every available measure, mission-critical. Mary Lou Rogers, one of the women who did this work, recalled that each component had to be inspected by three or four people before it was signed off, and that federal inspectors visited the plant regularly to check the work. The inspections were not, on examination, a sign that the work was untrusted. The inspections were the structural necessity of producing a piece of hardware that could not, once installed, be modified, and that human lives would depend on.

The “LOL” nickname and what it obscures

The engineers at MIT, including Margaret Hamilton, who led the flight software effort, sometimes referred to the women who wove the rope memory as “LOLs,” for “Little Old Ladies.” The nickname appears in Hamilton’s own remembered accounts, alongside other affectionately wry pieces of MIT lab vocabulary like “FLTs” for “Funny Little Things” (minor program bugs) and the “Auge Kugel” method (debugging by eyeballing the source code).

The nickname was, on the available evidence, used in some informal contexts within the program. Whether it was widely used at the time, or whether it became attached to the work in later retellings, is contested. Ken Shirriff, who has documented the AGC in considerable technical detail, has noted that he is not entirely convinced the term was current during the actual program. What is not contested is that the work itself was, by every available measure, skilled technical labor performed by trained employees of a major defense contractor under federal inspection, and that the affectionate nickname has tended, in the cultural register, to obscure the actual nature of what the women were doing.

The obscuring matters because the work, on close examination, was not the kind of work the cultural register tends to associate with the word “weaving.” The work required steadiness, attention, sustained concentration, and an error rate that approached zero. A single mistake in the wiring of a rope module would, in most cases, require starting the module over. The cost of a mistake, in time and material, was considerable. The economic value of the work was, accordingly, considerable. The wor