I'm following a proof I found online but I get lost halfway through
Suppose $Ax = b$ has more than one solution.
Let $x_1$ and $x_2$ be two possible solutions where $x_1 \neq x_2$
$Ax_1 = Ax_2$
$A(x_1-x_2) = 0$
$x_1 + k(x_1 - x_2)$ is a solution for $k \in R$
Why is this a solution to the linear system? Why are they adding $k(x_1 - x_2)$ to $x_1$?
$Ax = A[x_1 + k(x_1 - x_2)]$
$= Ax_1 + kAx_1 - kAx_2$
$= Ab + kb - kb$
$= b$
Suppose that $Ax = b$. So, we have $x_1$ and $x_2$ as our proposed solutions, with $x_1 \neq x_2$. Then:
$Ax_1 = b$
$Ax_2 = b$
$Ax_1 - Ax_2 = b-b = 0$
$A(x_1-x_2) = 0$
Now, I claim that $x_1 + k(x_1-x_2)$ satisfies $Ax = b$ for any $k \in \mathbb{R}$. This would clearly show that there are infinitely many solutions, since I can pick any value of $k$ that I want and I'd still have a solution.
All that's left is to actually prove that it satisfies the given equation. We have:
$A[x_1+k(x_1-x_2)] = Ax_1 + A(k(x_1-x_2)) = b +k[A(x_1-x_2)] = b +k0 = b$
This proves the desired result. I hope this helps you understand the proof better. The idea is that if you take a linear combination of $x_1$ and $x_2$, then you can show that it does lead to infinitely many solutions.