My first computer had an 80 MB hard drive. I remember the sticker on the box clearly, and I remember the drive showing up as slightly less than that in Windows. I assumed there was something wrong with my machine. My uncle, who was more technical, said "computers round things differently" and moved on. That answer was vague enough to feel satisfying and precise enough to seem credible. It carried me for years.

The actual explanation is both simpler and more irritating than "computers round things differently." It comes down to two industries — hardware manufacturers and software developers — having independently decided that a kilobyte means different things, and neither of them technically being wrong, and nobody ever fixing it properly, and you being stuck with a number on your screen that will never match the number on the box.

The MB vs MiB distinction is the formal attempt to clean this up. It mostly failed in practice. But understanding it explains the hard drive thing, the RAM thing, the internet speed thing, and about a dozen other minor frustrations that computer users treat as an unsolvable mystery.

The Real Problem — Two Definitions, One Word

In the rest of the world, "kilo" means 1,000. A kilogram is 1,000 grams. A kilometre is 1,000 metres. A kilojoule is 1,000 joules. This is the metric system, and it's been consistent for over two centuries. Kilo is always 1,000. Mega is always 1,000,000. Giga is always 1,000,000,000.

Computers work in binary — everything is powers of 2. And 2 to the power of 10 is 1,024. That's close to 1,000, but not 1,000. Early computer engineers needed a word for "1,024 of something" and they borrowed "kilo" because 1,024 is approximately 1,000. It was a convenience shortcut that made rough sense at the time.

So from very early on in computing, "kilobyte" started meaning two different things depending on who was talking. Hardware manufacturers — particularly hard drive makers — used the standard metric definition: 1 KB = 1,000 bytes. But operating systems and programmers largely used the binary definition: 1 KB = 1,024 bytes. Both groups called it a kilobyte. Neither group put a disclaimer on the label.

The gap looks small at the kilobyte level — 1,000 vs 1,024 is only a 2.4% difference. But the percentage doesn't stay flat. It compounds at every level up. By the time you're at terabytes, the difference between the manufacturer's number and the operating system's number is almost 10%. That's where the 1 TB = 931 GB mystery comes from. It's not rounding. It's a definitional gap that grows with every order of magnitude.

Where 1024 Comes From, and Why It Stuck

Binary is base-2. Everything in a digital computer is ultimately a 1 or a 0. Memory addresses, storage blocks, file system structures — they all work in powers of 2 because that's what the underlying hardware is built around. 2¹ is 2. 2² is 4. 2³ is 8. Keep going: 2¹⁰ is 1,024.

1,024 is the natural unit size for binary systems. Memory chips were designed in blocks of 1,024 bytes because that fits cleanly into binary addressing. A 10-bit memory address can point to exactly 1,024 locations — not 1,000, not 1,100, exactly 1,024. So "1 kilobyte of RAM" meant 1,024 bytes in practice because that was the actual physical block size.

Hard drive manufacturers never adopted the binary convention because it didn't suit them. A hard drive isn't built around power-of-2 block sizes the way RAM is — it's a physical spinning disk with sectors and tracks, and the engineers calculating capacity used straight decimal math. So when they labeled a drive "1 GB," they genuinely meant 1,000,000,000 bytes, not 1,073,741,824 bytes.

This wasn't dishonesty. Both groups were using the word "gigabyte" in the way that made sense for what they were building. The problem is that a consumer buying a hard drive and plugging it into a computer is bridging the two worlds — and the gap shows up immediately in the first number they see on screen.

The MiB Fix That Nobody Actually Uses

In 1998, the International Electrotechnical Commission — the IEC — got tired of the ambiguity and proposed a clean solution. They created new names specifically for the binary units. Kibibyte (KiB) for 1,024 bytes. Mebibyte (MiB) for 1,048,576 bytes. Gibibyte (GiB) for 1,073,741,824 bytes. Tebibyte (TiB) for 1,099,511,627,776 bytes.

Under the IEC standard, "MB" would unambiguously mean 1,000,000 bytes — the metric definition — and "MiB" would unambiguously mean 1,048,576 bytes — the binary definition. Hard drive manufacturers could keep using MB. Operating systems could switch to MiB. Everyone would know exactly what each number meant. No more confusion.

In practice, almost nobody switched. The words "kibibyte" and "mebibyte" never entered everyday language. Linux adopted GiB labeling in some contexts. macOS switched to decimal (GB) in 2009, which actually made Apple drives appear to have slightly more space on screen than Windows showed for the same hardware. Windows still shows binary-calculated numbers but labels them GB rather than GiB. Most people have never heard the word "mebibyte" in their lives.

Why Your Hard Drive Always Shows Less Than Advertised

Now that the definition gap is clear, the math is straightforward. Take the number of bytes on your drive — exactly as the manufacturer measured them — and divide by the operating system's version of the unit to get what you'll see on screen.

The confusion gets worse because formatting your drive for use does take a small additional chunk — the file system itself (NTFS, exFAT, ext4) needs space to store its own tables and metadata. On large modern drives this is small, maybe 200–500 MB. On top of the definitional gap, it adds a bit more distance between the box number and the screen number. Some drives also ship with recovery partitions pre-installed by manufacturers, which genuinely are extra space that Windows can see but not use freely.

But the recovery partition and file system overhead together account for maybe 10–20 GB on a typical drive. The majority of the gap on a 1 TB drive — 69 GB total — is the unit definition mismatch, not any actual missing space.

Why RAM Works Differently From Storage

Here's where it gets slightly interesting. RAM is always sold and reported in binary-aligned sizes. You don't see 8,000 MB RAM sticks — you see 8,192 MB (8 GiB) or 16,384 MB (16 GiB) or 32,768 MB (32 GiB). These are all powers of 2. When your system specs say 16 GB RAM, it actually means 16 GiB in binary terms — 17,179,869,184 bytes.

This is because RAM chips are physically built in binary-sized blocks. The addressing circuits, the memory controllers, everything about how RAM works is natively binary. There's no incentive or practical way for a RAM manufacturer to produce a 16,000,000,000-byte stick — the chips don't come in those sizes. So RAM manufacturers never had the same temptation that hard drive makers had to use decimal.

The result is that when Windows says you have 16 GB RAM, it means something slightly different than when it says your hard drive is 931 GB. The RAM number is accurate in binary terms. The hard drive number is the manufacturer's decimal figure recalculated in binary. Same label, different underlying math.

The Mbps vs MBps Trap That Catches Everyone

Storage confusion is one thing. Internet speeds have their own separate trap, and this one catches even people who understand the MB vs MiB issue.

Internet speeds are measured in bits, not bytes. A megabit per second (Mbps) and a megabyte per second (MBps) are not the same thing. One megabyte is 8 megabits. So a "100 Mbps" internet connection can transfer at most 12.5 megabytes per second — not 100 megabytes per second.

When you're downloading a file and your browser shows "12.3 MB/s," that's consistent with a 100 Mbps connection running at full speed. But it's very easy to read your plan as "100 MB" and feel like your 12 MB/s download is broken. ISPs advertise in Mbps because the numbers are 8 times larger and sound more impressive. Your download manager shows MB/s because bytes are what files are measured in. Nobody thinks to say that the units are different.

There's a secondary layer to this: internet speed Mbps uses the decimal definition of mega (1,000,000 bits per second), while the MB/s your download manager shows may use either decimal or binary megabytes depending on the software. Usually it's decimal MB/s in browser download bars, which makes the bits-to-bytes conversion cleaner. But some download clients show MiB/s in binary. The units are in the label if you look carefully, but most people never look at units — they just see the number.

The Numbers Side by Side — Every Unit Decoded

The lowercase "b" vs uppercase "B" distinction is easy to miss in fonts that don't render them clearly. Mbps vs MBps is an 8x difference in actual transfer speed. It's a distinction that matters and one that the industry has never made effortless for consumers to spot.

What macOS did differently — and why it matters

Apple quietly switched macOS to decimal gigabytes in 2009 with Snow Leopard. Before that, Macs showed storage in binary GiB (mislabeled as GB) the same way Windows does. After the switch, the same 1 TB hard drive that shows 931 GB on Windows shows 1,000 GB on a Mac. The actual drive is identical. macOS just switched to the manufacturer's definition of the unit.

This led to a brief period of complaints from people switching from Mac to PC — their drive "shrank" by 7% when they booted Windows. It also led to some confusion in the other direction: people on Macs felt like they had more space than they did, because their system was reporting in a unit that matched the box but didn't match binary file sizes in the same way. Neither approach is wrong. They're just different answers to the same unit ambiguity.