Table of Contents (click to expand)
No: even 200,000 chimpanzees typing one key every second until the universe dies would almost never reproduce a single book, so the classic “monkeys will eventually type Shakespeare” idea is true in theory and useless in reality. A hard drive makes it worse: a household magnet can wipe a drive, but its field is the same across the whole disk, so it cannot flip one microscopic bit at a time the way the drive's own write head does. And the bits are built to resist flipping on their own, which is exactly what a lucky random Windows install would need them to stop doing.
There is a scene in Breaking Bad where the characters park a truck full of powerful magnets next to a police evidence room. They switch the magnets on. A laptop inside is wiped clean through the wall. No one touches it. The data is just gone.
The plan works. It is also the last time in this article that a magnet does anything useful to a hard drive.
Because there is an old idea that keeps getting dressed up in new clothes. Give enough monkeys enough typewriters and enough time, and one of them eventually bangs out the complete works of Shakespeare. So here is the modern version. Give a million monkeys a million hard drives and a million strong magnets, and let them wave the magnets around forever. Does one drive eventually end up holding a clean, bootable copy of Windows?
The intuitive answer is "sure, eventually, that's just what the theorem says." That answer is wrong twice, for two reasons from completely different parts of physics.
What Is The Infinite Monkey Theorem?
Imagine a monkey mashing keys on a keyboard at random, every key equally likely. Given an infinite amount of time, the theorem says, the monkey will eventually type any text you like, including all of Shakespeare, purely by chance.
The math behind that is honest. If something has any chance above zero of happening on a single try, then over infinitely many tries it happens with certainty. Hold on to that word "eventually," though. It is doing an enormous amount of hidden work.
Here is the part almost nobody knows. The monkeys were never really about monkeys. The French mathematician Émile Borel introduced them in 1913, while writing about statistical mechanics, the physics of why heat flows one way and never runs backward. Borel wanted a picture for something so unlikely it would never once be seen: a million monkeys reproducing every book in the richest libraries of the world. So the monkeys started life as a ruler for the impossible, not a promise of the possible. Keep that in your pocket.
What Happens When You Take The Infinity Away?
The pure theorem is a proven piece of math. The interesting question is what happens when you take the infinity away, because the real universe does not have one to spare. Two mathematicians recently ran exactly that calculation, with numbers grounded in reality. Not infinite monkeys, but 200,000, roughly the chimpanzee population alive today. A 30-key keyboard. One keystroke per second. And a clock running not forever but until the heat death of the universe, around a 1 followed by 100 zeros in years.
Take the smallest possible target, the word "bananas." With 30 keys, each press has a 1 in 30 chance of being the letter you want, and every extra letter you need in a row multiplies the odds against you by 30 again. "Bananas" is seven letters, so all seven in a row comes to about 1 in 22 billion per attempt. A chimp pressing one key a second for a 30-year working life gets just under a billion presses, which works out to about a 5% chance of typing the word within its own lifetime.
So a chimp will, given a career, quite possibly type "bananas." It is the most on-brand result in all of probability theory. Now stretch the target. Take Curious George, a children's book of around 1,800 words. The odds of a chimp reproducing it come out as a decimal point followed by roughly 15,000 zeros before the first digit. Even all 200,000 chimps typing until the heat death of the universe do not make a dent. The mathematicians' verdict on the classic theorem: somewhat misleading. True in principle. Useless in a universe with a shelf life.

Now Swap The Typewriter For A Hard Drive
The typewriter version is hopeless. The hard-drive version is hopeless in a more expensive way.
A hard drive does not store letters. It stores bits. Each bit is a tiny patch on a spinning platter, magnetized one way for a zero and the other way for a one. A one-terabyte drive, as sold, holds a trillion bytes, or 8 trillion bits, each of which can be a zero or a one independently.
So the number of possible things a single drive could contain is 2 multiplied by itself 8 trillion times. Written out, that is a 1 followed by roughly 2.4 trillion zeros. For comparison, the whole observable universe holds something like a 1 with 80 zeros' worth of atoms. The drive's list of possible contents buries that so far down that "astronomically larger" is a wild understatement.

And "working Windows" is a far smaller target than it sounds. A bootable Windows drive is not one lucky pattern. It is a stack of requirements that all have to be right at once: a valid boot sector, a consistent master index of every file, a coherent file tree tens of gigabytes in size, program files whose internal headers parse, and driver signatures that verify. One wrong byte in the boot sector and you do not get slightly-worse Windows. You get a brick.
Even if the number of valid bootable Windows drives were somehow a 1 with 10 billion zeros, an absurdly generous count, it would still be a rounding error inside a space of a 1 with 2.4 trillion zeros. The odds are, for every practical purpose, indistinguishable from zero. This is where the analogy should break. It actually breaks earlier, for a reason the probability never gets to.
Why A Magnet Can't Write To A Hard Drive
Here is the reason nobody in the original story reached, because they never got past the counting. A magnet cannot write to a hard drive at all. Not "unlikely to write something useful." It cannot flip one chosen bit. Ever.
To see why, look at how a drive actually writes. The data sits in a thin magnetic coating made of tiny grains, each less than ten nanometers wide (a nanometer is a billionth of a meter), and one bit is a patch of many such grains. Those grains are built to be stubborn on purpose. It takes a strong magnetic field to flip one, far stronger than a fridge magnet. Engineers call that stubbornness coercivity, and drive coatings are made with plenty of it, so they hold firmly whatever was written.
The write head manages it with one trick: it squeezes its magnetic field down onto a single patch. The tip of the head hovers just a few nanometers above the spinning disk. That close, the field is fierce directly under the tip. A hair's width to either side, it has already faded to almost nothing. So the head can flip one patch of grains and leave its neighbor, right beside it, completely alone. It is a scalpel.
A handheld magnet is the exact opposite. Its field is essentially uniform across a whole platter, so it has no fine control at all. It flips nothing, or it flips everything in a region at once. That is why a strong enough magnet can erase a drive, as the Breaking Bad magnet scene gets basically right. But flipping one specific bit to one specific value, over and over, to write a file? A magnet is not a scalpel. It is a house fire.

Why The Bits Won't Even Stay Where You Put Them
Suppose we take pity on the monkeys and hand them a magic flipper that can flip any bit they like. The physics still fights back, from the other direction.
A bit keeps its value because there is a kind of wall holding its grains pointing one way. How high that wall is depends on just two things: how stubborn the material is, and how big the grain is. Shrink the grain too far and the wall gets so low that the ordinary warmth of the room, the endless tiny jiggling of everything at room temperature, is enough to flip the grain on its own. No magnet needed. Engineers call this the superparamagnetic limit, and drives are built to stay well clear of it, so a bit survives about ten years without one accidental flip.
Sit with what that means. The coating is engineered, atom by atom, to resist random flipping, because random flipping is exactly what destroys your data. The newest drives make this even worse. Their coating is so stubborn it can barely be written at room temperature at all, so a tiny laser has to warm each patch first, just for an instant, before the head can change it.
So, could monkeys with magnets ever produce Windows? No. And the two reasons it fails turn out to be the same reason, which is the quietly beautiful part. The probability argument says the target is too small to ever hit in the lifetime of the universe. The physics argument says a magnet cannot aim at the target anyway, and that the coating is built to undo any random hit. Both come straight from statistical mechanics, the physics of how enormous numbers of tiny particles behave, which is exactly the field Borel invented the monkeys to illustrate.
His monkeys were never a claim that monkeys could write books. They were a ruler for "so improbable it will never be observed," built to show that the one-way flow of heat we met earlier is, for all practical purposes, inevitable. The hard-drive version proves his point better than the typewriter ever did: thermal jiggling and overwhelming odds are the very same statistical mechanics the monkeys were created to make you feel. The monkeys came home.
They keep excellent company, too. There is the Boltzmann brain, a fully formed mind that fluctuates into existence out of random noise, and Maxwell's demon, the tiny sorter that would break thermodynamics if only it could exist. The monkeys just can't type.
References (click to expand)
- A numerical evaluation of the Finite Monkeys Theorem — Woodcock & Falletta, Franklin Open (2024)
- It would take longer than the life of the universe for a monkey to randomly type Shakespeare — University of Technology Sydney, via ScienceDaily
- Warren-Crow, H. Shakespeare and Nonhuman Intelligence, Cambridge University Press — on Borel’s 1913 paper and Le Hasard (1914) as the theorem’s origin
- Directed self-assembly of block copolymers for use in bit patterned media fabrication — Griffiths et al., J. Phys. D: Appl. Phys. 46, 503001 (2013)
- Perpendicular Magnetic Recording Technology (white paper) — Western Digital
- Ultrasensitive mechanical detection of magnetic moment using a commercial disk drive write head — Tao et al., Nature Communications 7:12714 (2016)
- Thermal effect limits in ultrahigh-density magnetic recording — Weller & Moser, IEEE Transactions on Magnetics 35(6), 4423 (1999)
- Master File Table — Microsoft Learn
- Windows 11 Specifications (minimum 64 GB storage) — Microsoft







