We use the null space for the proof, therefore, I will explain the null space a bit. If you know about the null space, of course you can skip this entry.
The null space of a matrix A is a set of non-zero vector x that satisfies Ax = 0.
Let me show you another example, but this time a rectangle matrix A that has null space.
By the way, these are all about definition of null space. I could finish the story about null space at this point, however, I would like to think about the meaning of null space a bit more. I would like to talk about the existence of the inverse matrix, so let's talk about square matrices. Because a rectangle matrix can be interpreted as a matrix that has too many equations or too less equations, there is no inverse for such matrices.
null space is related with a matrix is singular or not, that means the matrix has independent columns or not. Because, if we write down an A with column vector, and multiply with x,
Therefore, Ax is a linear combination of the column vectors a_i with coefficients x_i. If this result is 0, x is a null space. This is one aspect of null space.
When we look at the columns of Ax = b, b is the linear combination of column vector of A. A space consists of column vectors is called column space. If we have n independent columns in n-dimensional space, we could express any point in the n-dimensional space. This means we have a solution for any b and we have the inverse. If A has the inverse, the linear combination of the column vector becomes 0 iff x= 0. This also means there is no null space (= or only 0 exists).
Now we see the null space, column space, and the existence of the inverse. Do you think you have now a bit clearer view about the relationship among them?
Thanks to Marc D. who listened my explanation and gave me comments.