What Is Computation?

Computation takes information, applies rules or operations, and produces a result. Understand it through simple examples and its connection to computers and AI.

Illustration for the current article

Computation takes information, applies rules or operations, and produces a result. A calculator adding two numbers is computing, but so is software sorting a customer list or finding a route on a map. Calculation is one kind of computation—not the whole idea. The useful pattern is: input, processing, result. To understand what a computer is doing, ask what information it receives, what operations it performs, and what result it produces.

A Simple Example of Computation

Imagine a customer list containing Meera, Arun, and Zoya. You want the names in alphabetical order. The input is the original list. The rule is to compare names alphabetically and put earlier names before later ones. The result is Arun, Meera, Zoya.


Nothing has been added up. Yet computation has taken place: information was processed according to a rule to produce a result. The same idea applies whether you sort three names by hand or software sorts a much larger list.

Computation Is Broader Than Calculation

Adding an invoice total is calculation, and calculation is one kind of computation. Sorting customer names, resizing a photograph, and finding a route on a map are also computations. Each takes information, performs operations, and produces a result.


The result does not have to be a number. It can be an ordered list, a modified image, a route, or text. What matters is the process applied to the information.

Algorithm, Program, and Computation: What Is the Difference?

An algorithm describes a step-by-step method for solving a problem. For our customer list, it could specify how to compare names and rearrange them until they are in order. A program expresses instructions in a form a computer can execute. Computation is what happens as those operations are carried out on information.


Think of the distinction as method, instructions, and execution. The algorithm describes how to sort; the program implements that method; running it on your list performs the computation.

How Computers Carry Out Computation

A digital computer represents information using bits: values written as 0 and 1. Physical circuits operate on signals representing those bits. Logic gates implement simple operations, and larger circuits combine them to carry out instructions.


A processor executes program instructions using data held in memory. In the sorting example, the computer represents the names, compares them according to the chosen ordering rules, and stores the reordered list. Bits do not form gates; circuits implement gates that operate on signals representing bits.


For the representation step, read how binary represents numbers and letters. For the machine itself, read what a computer really is. To explore executable instructions, read what a program is. For a closer look at circuits, continue with NAND to Intelligence.


NAND gates can be combined to express any Boolean function, a property called functional completeness. That is different from saying an individual gate is a universal programmable computer.

Does Computation Require an Electronic Computer?

No. People can carry out computations by following procedures on paper, and mechanical devices can perform computations too. Before electronic computers, “computer” was also a job title for people employed to perform calculations.


The distinction is useful: computation is the process, while a computer is a system that carries it out. Electronic hardware makes many computations fast and repeatable, but electronics is not part of the basic definition.

What Computation Cannot Do

Theoretical limits

Some precisely defined problems have no algorithm that can correctly solve every possible case. For example, there is no general algorithm that can decide, for every program and input, whether that program will eventually stop. This is the halting problem. It does not mean we can never determine whether a particular program stops.

Practical limits

Other problems can be solved in principle, but a particular computation may need more time or memory than we can provide. A method that works for a small input may become impractical as the input grows. Faster hardware can help with practical limits; it does not remove a proof that no general algorithm exists.

How AI Uses Computation

AI uses computation both when a model is trained and when it is used. During training, computations use data and an objective to adjust model parameters. During inference, the trained model processes an input and produces an output using those parameters.


Learning does not mean escaping computation. The training procedure itself performs operations, while the learned parameters influence what the model produces. AI progress depends on data, algorithms, model design, hardware, and engineering—not computation at scale alone.


Producing a fluent answer is not, by itself, evidence of human-like understanding. A useful output can still contain mistakes or miss important context. People remain responsible for judging whether an answer is appropriate and reliable.


Continue with How AI Learns From Data: A Simple Non-Technical Guide for a closer look at training.

Key Takeaway

Computation takes information, applies rules or operations, and produces a result. When you encounter a digital tool, ask three questions: What information goes in? What operations are performed? What result comes out? That gives you a starting point for understanding the tool without treating it as magic.

Frequently Asked Questions

What is computation in simple words?

Computation is the process of taking information, applying rules or operations, and producing a result.

Is computation the same as calculation?

Calculation is one type of computation. Computation also includes tasks such as sorting a list, processing an image, or finding a route.

What is the difference between an algorithm and computation?

An algorithm describes a method. Computation is the process of carrying out operations on information using that method.

Does computation always produce the same answer?

A deterministic computation produces the same result when its relevant inputs and starting state are the same. Other computations can include randomness or changing inputs.

How does AI use computation?

AI uses computation during training to adjust model parameters and during use to process inputs and produce outputs. Those outputs still require appropriate human evaluation.

KEY TAKEAWAY

Before you move on, pause with the central idea from this lesson and connect it to the next step in your learning path.