Summary: What vectors actually are

The word “vector” means an arrow in physics, a list of numbers in code, and “an element of a vector space” in a math textbook, and people quietly suspect those are three different things. They are not. They are one object seen from three angles, defined by just two operations: you can add two vectors, and you can scale one by a number. Get that solid and the rest of linear algebra has something to stand on, because the vector is the atom every later topic is built from. This is the scan-it-in-five-minutes version.

Core ideas

• A vector has three faces. The physics view is an arrow: a displacement with a length (how far) and a direction (which way), the same vector wherever you draw it. The computer science view is an ordered list of numbers, where the order matters and the length of the list is the dimension. The math view is the most general: a vector is anything you can add and scale coherently, which is why functions and polynomials count too.
• The textbook definition feels empty on purpose. “An element of a vector space” refuses to commit to arrows or lists; what it names is the one thing both share, the two operations. That is the punchline of the lesson, not a dodge.
• A coordinate system is the bridge: lay down axes and a unit length, and every arrow gets a unique list of numbers, and every list becomes a unique arrow. Numbers for the computer, arrows for your head, the same object underneath. You spend the rest of the series crossing this bridge.
• A coordinate (or component) is one entry in the list, saying how far to move along one axis; the dimension is how many there are. Crucially, the coordinates are a description of the vector in a chosen frame, not the vector itself: change the axes and the numbers change while the vector stays put.
• Vector addition is component by component, [1, 2] + [3, 1] = [4, 3], and geometrically it is tip-to-tail: walk the first arrow, then the second from its tip, and the endpoint is the sum. It is not list concatenation; [1, 2] + [3, 4] is [4, 6], never [1, 2, 3, 4].
• Scalar multiplication multiplies every component by one number. A scalar stretches the arrow (factor above 1), squishes it (below 1), or flips it (negative), without rotating it off its line. Addition and scaling are the complete toolkit; spans, matrices, and every later operation are assembled from just these two moves.
• This is why the math is worth your time for AI. Inside a model, words are vectors and the internal state is vectors being added and scaled; the famous “king minus man plus woman lands near queen” is exactly the component arithmetic you just learned, run in a few hundred dimensions instead of two. High dimensions are not mystical: a 300-dimensional vector is a list of 300 numbers, and every rule from the flat grid applies unchanged.
• Pitfalls worth naming: thinking the arrow must start at a fixed spot (it is a displacement, rooted at the origin only by convention); thinking the coordinates are the vector (they are one frame’s description); treating addition as concatenation; and believing a vector must be an arrow or a list (the general definition is “anything you can add and scale”).

What changes for you

Before this lesson, “vector” was a word that shifted meaning depending on who was talking, so it never quite settled into anything. Now it points at one object you can hold from three sides, with a clear rule for moving between them. When you next meet matrix notation in a paper or a library doc, you have a sharper move available: read the numbers as arrows in space, and ask what operation is being done to them, rather than treating the symbols as a wall. Every later lesson in this track (spans, transformations, the determinant, eigenvectors) is built on the two operations you just pinned down, so the work you did here is the work that makes the rest readable.