What counts as intelligent?
In 1950 Alan Turing proposed judging a machine by its behaviour: if you can't tell it from a person in conversation, call it intelligent. In 1956 the Dartmouth workshop named the field artificial intelligence. Modern AI, as Russell and Norvig's textbook frames it, asks something more practical: can we build an agent that acts rationally — that does the right thing given what it knows?
This course builds the classic ideas one at a time, in JavaScript you write and test: agents, search, games, constraints, logic, probability and learning. The large language models you've heard about sit on top of these same foundations — a chess engine, a route planner, a spam filter and a neural network are all agents of one kind or another.
Percepts, actions and agent functions
An agent perceives its environment through sensors and acts on it through actuators. What it senses at one moment is a percept. An agent is defined by its agent function: what action it takes for any sequence of percepts. The simplest kind, a simple reflex agent, looks only at the current percept and applies condition–action rules: if the square is dirty, suck.
function reflexVacuum([location, status]) {
if (status === "Dirty") return "Suck";
return location === "A" ? "Right" : "Left";
}A thermostat: Too cold → heat on. Warm enough → heat off. A thermostat is a reflex agent: no memory, no plan, one rule per situation — and for its tiny world that's perfectly rational.