Everyone Means Something Different
Say the word "robot" and five people picture five different things. A thermostat that turns the heat on. A drone hovering over a field. A chatbot answering questions. A giant orange arm welding car doors. A vacuum bumping around the living room. They cannot all be the same kind of thing — and yet we reach for the same word. So before we can study robots, we need a sharper question: what actually separates a robot from a fancy appliance, or from a clever piece of software?
There is no single official answer that everyone agrees on, but the field of robotics has a useful working definition. The rest of this guide builds it up piece by piece, then shows why the messy physical world is what makes robots genuinely hard.
A Working Definition
Here is the line most roboticists draw. A robot is a physical machine that senses its surroundings, decides what to do, and then acts on the world through its own body — and it does this loop on its own, without a human steering every move. Hold that sentence up against the confusing examples and the fog starts to clear.
A chatbot is out: it has no body and never touches the physical world — it lives entirely in software. A factory machine that repeats one fixed motion forever is out too, but for the opposite reason: it does not really sense or decide, it just replays a script. That kind of blind repetition is automation, not robotics. A robot sits in the sweet spot between them — it has a body like the appliance, and it makes choices like the software.
The Three Ingredients
The definition hides three ingredients, and a machine needs all three to earn the name. Let's follow a robot vacuum cleaning your floor and watch each one show up.
- Sensing — taking in the world. The vacuum has bump sensors on its rim, a downward-facing cliff sensor that catches the edge of a staircase before it drives off, and often a camera or a spinning laser that measures distances to the walls. This is how raw reality gets turned into numbers it can work with.
- Computation and decision — making sense of it. A small computer inside builds a rough map, figures out where it has already cleaned, notices "wall ahead," and chooses to turn rather than keep ramming. This is the deciding step — the difference between reacting blindly and choosing.
- Actuation — pushing back on the world. Motors spin the wheels to drive forward and pivot; another motor spins the brush; a fan creates suction. An actuator is any part that converts a decision into real physical motion or force. No actuators, no action — and a machine that cannot act on the world is not a robot.
Now the washing machine puzzle solves itself. It does sense water level and temperature, and it does act by spinning a drum. What it lacks is the middle ingredient with any real autonomy: it runs a fixed timed program and makes no genuine choices about an unpredictable world. The vacuum, by contrast, faces a different room every day and must decide its own path. That third ingredient — deciding for itself in a situation nobody scripted — is what tips it over the line.
Why a Body Changes Everything
Having a body in the physical world is not a minor detail — it is the thing that makes robotics its own field, and brutally hard. This idea has a name: embodiment. A chatbot that gives a wrong answer can simply try again; nothing breaks. A robot arm that misjudges by two centimeters knocks a glass off the table, and there is no undo. The world does not wait, does not reset, and never repeats itself exactly.
A body also means dealing with everything physics throws at you: sensors that report slightly wrong numbers, motors that overshoot, friction, gravity, a floor that is dustier in one corner, lighting that changes by the hour. A robot must keep working despite all of it, in real time, with real consequences. That is why a task trivial for a person — picking up a soft towel, walking across gravel — can be far harder for a robot than tasks we consider "intelligent," like playing chess. (That surprising reversal even has a name, Moravec's paradox, which a later guide unpacks.)
The Loop That Ties It Together
Notice that the three ingredients are not a one-time checklist — they run over and over, many times a second. Sense the world, decide what to do, act, then sense again to see what changed, and around once more. That endless cycle is the heartbeat of every robot, and it has a name: the sense–plan–act loop. It is the single most important idea in this whole track, and the very next guide takes it apart step by step.
Robots also come in wildly different shapes — rolling, walking, flying, swimming, arms bolted to a factory floor — and there is an orderly way to sort them, called a robot taxonomy. But every one of them, no matter the shape, lives by the same three-ingredient loop. Get that loop in your head and you will understand the skeleton of any robot you ever meet.