What a Computer Actually Does: CPU, RAM, Disk, OS

A computer is a very fast, very stupid machine that follows instructions

Computers feel like magic, and that feeling is a problem — it makes everything downstream (coding, why things are slow, why the internet works the way it does) feel like magic too, so you never get a grip on any of it. Strip the magic away and a computer is almost embarrassingly simple in concept: it's a machine that does billions of tiny, dumb operations per second, and follows instructions exactly — including your mistakes. It has no understanding and no judgement. Everything it does — every app, game, video call — is those tiny operations, stacked unimaginably high and run unimaginably fast.

Once you can picture the four main parts and how they pass work between them, a lot stops being mysterious: why closing tabs speeds things up, why more RAM helps, why an SSD made your old laptop feel new, what "loading" is actually doing. This is the foundation the whole Technical column stands on.

What you'll have at the end

• The four parts — CPU, RAM, disk, OS — and what each is for
• The speed hierarchy that explains most "why is it slow?" questions
• A clear model of what happens when you open an app

Everything is instructions on numbers

At the bottom, a computer only does two kinds of thing: store numbers and do simple operations on them (add, compare, move this number there). Everything else is built from that. A photo is numbers (the colour of each pixel). This text is numbers (a code per character). A program is numbers too — a list of instructions. The machine reads an instruction, does it, reads the next, forever, billions of times a second. That's the whole game; the sophistication is entirely in how many and how fast, not in any single step being clever.

And it's all binary — every one of those numbers is stored as bits (0s and 1s, because a switch is easiest to build as on/off). Eight bits make a byte; bytes make kilobytes, megabytes, gigabytes. When you hear "16 GB of RAM" or "512 GB SSD", that's just how many bytes it can hold.

The four parts and what each is for

CPU — the worker

The Central Processing Unit is the bit that actually does the instructions — the worker. It runs the fetch → decode → execute loop billions of times a second (its "clock speed", measured in GHz, is roughly how many steps per second). Modern CPUs have several cores, meaning several workers doing instructions in parallel — which is why more cores help when software can split work up. The CPU is fast; almost everything else is about keeping it fed.

RAM — the desk

RAM (Random Access Memory) is the working space — think of a desk. It's where the programs and data you're using right now sit, because the CPU can reach it very fast. Two defining traits: it's fast, and it's volatile — it forgets everything when the power goes off. So RAM holds your open apps and open files while you use them; it is not where things are kept. More RAM = a bigger desk = more you can have open at once before the machine struggles.

Disk — the filing cabinet

The disk (an SSD in most modern machines, or an older spinning hard drive) is permanent storage — the filing cabinet. It keeps your files, apps and the operating system even when powered off. It holds far more than RAM, but it's much slower to reach. So things live on disk when not in use, and get loaded into RAM (onto the desk) when you actually open them. An SSD is dramatically faster than an old spinning drive, which is why swapping to one makes an old computer feel new — less waiting on the slow part.

OS — the manager

The Operating System (Windows, macOS, Linux, Android, iOS) is the manager that runs the whole office. It decides which program the CPU works on and when (so many apps appear to run at once), hands out RAM to programs, organises files on disk, and talks to the hardware (screen, keyboard, wifi) so individual apps don't have to. Every app runs on top of the OS and asks it for what it needs.

The speed hierarchy (this explains most "why slow?" questions)

The single most useful idea here: the parts run at wildly different speeds, and performance is mostly about the fast parts waiting on the slow ones.

CPU registers/cache  — fastest (tiny, on the chip)
RAM (the desk)       — fast, but bigger
SSD/disk (cabinet)   — much slower than RAM
network (internet)   — slowest of all

Each step down is enormously slower than the one above. This one hierarchy explains a lot:

• Why more RAM helps: if your open apps don't fit on the desk (RAM), the OS shuffles data back to the slow disk ("swapping"), and everything crawls. A bigger desk means fewer trips to the cabinet.
• Why closing tabs/apps speeds things up: each one holds desk space (RAM); free the desk, less swapping.
• Why an SSD transformed your old laptop: loading and swapping hit the disk constantly; making the disk far faster removes the wait.
• Why "loading" exists: it's data being moved from the slow disk into fast RAM so the CPU can work on it.

What happens when you open an app (put it together)

1. You double-click. The OS finds the app's instructions on the disk.
2. It loads them from disk into RAM — the slow "opening…" step, moving from cabinet to desk.
3. The CPU starts executing those instructions from RAM, billions per second.
4. As you use the app, its data lives in RAM; when you hit Save, it's written back to the disk so it survives. Close without saving and the desk is cleared — it was never in the cabinet.

That's it. Four parts, passing work along a speed hierarchy, managed by the OS. No magic — just a lot of simple steps, very fast.

What this means for you

• A computer is a fast, literal instruction-follower on numbers (bits/bytes) — no understanding, no magic.
• CPU = the worker (does instructions); RAM = the desk (fast, temporary workspace); disk = the cabinet (slow, permanent storage); OS = the manager.
• The speed hierarchy (CPU ▸ RAM ▸ disk ▸ network, each far slower) explains most performance: keeping the fast CPU fed is the whole battle.
• More RAM / closing apps / an SSD all help by reducing waiting on the slow parts.
• "Loading" = moving data from slow disk to fast RAM. Save = writing from RAM back to disk.

Exercise (25 min, verifiable output)

1. Find your specs: how much RAM, what kind of disk (SSD or HDD) and how big, how many CPU cores. (System settings / About this device.)
2. Open your task/activity manager (Task Manager / Activity Monitor) and watch CPU and RAM usage as you open and close a few apps. Note what happens to RAM when you open a heavy app or many tabs.
3. Draw the four parts and label what each does.
4. Write the open-an-app sequence in your own words: disk → RAM → CPU → (save) → disk.
5. Explain, in one line each, why more RAM helps and why an SSD is faster than an old hard drive.

✅ Finish check: your machine's specs identified, a labelled CPU/RAM/disk/OS model, the open-an-app sequence written in your words, and the RAM/SSD "why" explained.

Summary card

• Computers do billions of tiny, dumb operations per second on numbers (bits → bytes). No magic.
• CPU = worker (executes instructions) · RAM = desk (fast, volatile workspace) · disk/SSD = cabinet (slow, permanent) · OS = manager.
• Speed hierarchy: CPU ▸ RAM ▸ disk ▸ network — each far slower; performance = keeping the CPU fed.
• More RAM, fewer open apps, an SSD = less waiting on slow parts.
• Open app: disk → load into RAM → CPU runs it → Save writes back to disk.

Sources

1. Petzold, C. — Code: The Hidden Language of Computer Hardware and Software, 1999
2. Patterson, D. & Hennessy, J. — Computer Organization and Design, 1994
3. Tanenbaum, A. — Modern Operating Systems, 1992

Next lesson: TC-02 — How the Internet Works: DNS, HTTP, Servers, Browsers (L1) Related: TC-02 How the Internet Works · TC-03 The Terminal · TC-11 Security Basics · AI-01 What a Language Model Does Path: Ship Your First Thing — 1/7