Tech & AI, Explained

what-is-a-computer

8 min read

What is a computer?

Status: 🟩 COMPLETE Last updated: 2026-06-20 Plain-English tagline: A machine that follows instructions step by step, very fast — and that’s it. Everything else is a layer built on top of that one idea.

In plain English

A computer is a machine that does one thing: it follows instructions, one after another, at enormous speed. The instructions are tiny — “add these two numbers,” “compare these values,” “store this in memory,” “send this signal to the screen.” Billions of those tiny instructions per second add up to everything you see a computer do.

Modern computers feel magical because the instructions happen so fast and we never see them. When you click a button on a website, hundreds of millions of little instructions ran between the click and the screen updating — and they all happen in a fraction of a second, so the response looks instant.

There is no “computer brain” understanding anything. There is no consciousness, no intent. It’s just a metronome of instructions, ticking away at 3 billion times per second.

Why it matters

Every other concept in this encyclopedia builds on this idea. When we say “JavaScript runs in the browser,” what we mean is: the browser hands the computer a list of JavaScript instructions, the computer follows them step by step, and the result is the things you see on the page. Same with everything else — databases, AI models, the internet itself. It’s all instructions following each other.

If you grasp this single thing, software stops being magic and starts being plausible. You’ll never again wonder “but how does the computer understand what I typed?” — it doesn’t understand. It follows instructions, very fast.

The four parts every computer has

Every computer — phone, laptop, server, smart fridge — has four functional parts:

PartWhat it doesReal-world analogy
InputTakes information from the outside worldYour hands writing on paper
ProcessingRuns the instructions, makes decisionsYour brain thinking
Memory & storageHolds information for laterA desk (short-term) + a filing cabinet (long-term)
OutputShows the resultReading aloud what you wrote

Input

Keyboard, mouse, touchscreen, microphone, camera, network — anything bringing information in. Even an internet connection is “input” from this perspective: data arriving from elsewhere.

Processing — the CPU

The CPU (Central Processing Unit) is the part that actually runs the instructions. It’s a small silicon chip with billions of transistors (tiny electrical switches). The CPU has a “clock” — it ticks at, typically, 3–5 GHz (3 to 5 billion ticks per second). Each tick, the CPU can do one or two tiny operations. That’s where the speed comes from.

Modern CPUs have multiple cores — separate processing units inside one chip. A 4-core CPU can do four things at once. A 16-core server CPU can do sixteen.

Memory — RAM

RAM (Random Access Memory) is where the computer keeps things it’s working on right now. When you open a program, the program gets loaded from storage into RAM. When you close it, RAM is freed up. RAM is fast but temporary: turn off the power and everything in RAM disappears.

Modern computers have 8–64 GB of RAM. Servers can have hundreds of gigabytes.

Storage — SSDs and hard drives

Storage is where information lives permanently — your photos, documents, the operating system itself. It’s slower than RAM but it survives power-off. Modern computers use SSDs (Solid-State Drives) which are faster and more reliable than the spinning-platter hard drives they replaced.

A typical laptop has 256GB–2TB of SSD storage.

Output

Screen, speakers, network out, printer, the LEDs on a server’s front panel — anything sending information back to the world.

Hardware vs software

A useful split:

  • Hardware is the physical stuff: the chips, the cables, the screen, the keyboard. You can hold hardware in your hand (or drop it on your foot).
  • Software is the instructions the hardware follows. Software has no physical form — it’s patterns of electrical states in memory and on disk. You can copy software perfectly; you can’t copy hardware.

Software is built in layers, each one running on top of the one below:

LayerExamplesWhat it does
FirmwareBIOS, UEFIThe most basic instructions burned into the hardware itself — what runs when you press the power button
Operating system (OS)Windows, macOS, Linux, iOS, AndroidThe boss software — manages memory, schedules programs, talks to hardware
System librariesVariousShared code that programs use to talk to the OS
ApplicationsBrowser, VS Code, Spotify, your Next.js appThe programs you actually interact with

When you run a Next.js app on Vercel, the layers stacked under it are: the Vercel platform → a container running Linux → Node.js → your Next.js code → your app’s logic. Each layer relies on the layers below it.

How a computer actually “thinks”

It doesn’t. But here’s how the metaphor breaks down into something concrete.

The CPU has a very small list of basic instructions it can perform — typically a few hundred different ones. They look like:

  • ADD: take two numbers, add them
  • LOAD: fetch a value from memory
  • STORE: put a value into memory
  • COMPARE: check if two values are equal
  • JUMP: go to a different instruction in the list

Everything a computer does — playing video, running ChatGPT, rendering a webpage — is composed of these tiny operations. The CPU just runs them, one after another, at billions per second. Patterns of these operations form algorithms; collections of algorithms form programs; layers of programs form your laptop’s experience.

The pattern is similar to a chef following a recipe — but the recipe has been compiled down to instructions like “lift right arm 2 cm” and “rotate wrist 30 degrees,” and the chef does a billion of those per second.

What’s special about computers

Three properties make computers more powerful than any other machine humans have built:

  1. They follow ANY instructions. A toaster does one thing. A car does one thing. A computer does whatever you give it instructions to do — graph data, play chess, identify a face, compose music. The same physical machine performs millions of different tasks based purely on which instructions it’s running.

  2. They’re absurdly fast. A modern CPU does ~10 billion simple operations per second. A modern GPU (graphics processor — also used for AI) does trillions. To match this, every human on Earth would have to do one simple math problem per second, in unison.

  3. They never get bored or tired. A computer runs the same loop 10 billion times in a row without getting impatient. This is what makes them useful for tasks humans find tedious but that benefit from being done very many times — like rendering pixels, sorting lists, or checking spam.

A concrete example: what happens when you open a browser tab

  1. Input: You click the browser icon.
  2. The OS notices the click, finds the browser program on storage, loads it into RAM.
  3. The CPU starts executing the browser’s instructions. First instruction: “draw a window.” Next: “ask the user for a URL.” Etc.
  4. You type a URL. Each keystroke is input; the CPU runs the instruction “display this character,” “compare to known URLs for autocomplete,” etc.
  5. You press Enter. The browser’s instructions say “send a network request.” The CPU passes that off to the OS, which talks to the network card.
  6. The network card sends the request out into the world, eventually reaching a server.
  7. The server runs its own instructions to produce a response, and sends it back.
  8. Your browser receives the response, runs HTML/CSS/JS instructions to figure out how to display it, and the CPU sends instructions to the screen to draw the page.

Steps 1–8 take roughly half a second. In that half-second, hundreds of millions of CPU instructions ran across your laptop and the server.

Common gotchas

  • “My computer is slow” usually isn’t the CPU. It’s almost always one of: not enough RAM (programs swapping to disk), a full or failing SSD, network latency, or a misbehaving piece of software hogging resources. Replacing the CPU rarely helps; adding RAM or replacing the storage usually does.

  • Storage size and RAM size are completely separate things. “I have 1TB of storage so my computer is fast” — no. Storage and RAM are different layers. You can have a huge SSD and still feel slow if RAM is low.

  • A “64-bit computer” doesn’t mean 64x faster. It means the CPU’s basic word size is 64 bits — which mostly matters for how much memory it can address (up to ~16 exabytes, vs ~4GB on a 32-bit CPU). Almost every computer made in the last 15 years is 64-bit.

  • GPUs are not just for graphics anymore. A GPU is essentially “a CPU with thousands of small, parallel cores instead of a few big ones.” This turns out to be great for AI training and inference. NVIDIA’s stock price tripling in 2024 is the story of this realization.

  • “The cloud” still runs on physical computers. When you “deploy to the cloud,” your code is running on someone else’s physical server in a data center somewhere. The fact that you don’t see or manage that server is the whole value proposition — but the silicon is real.

See also

Sources

Graph View

Backlinks