In Trebla's answer, he has multiplied both sides of the assumption by (1+x) which will always be positive because x ≥ 0, so you can do it!
In my answer, it's a matter of deduction and calculation.
I have used the assumption [(1+x)^k - 1 - kx] and multiplied it by (1+x) also. So:
(1+x)[(1+x)^k - 1 - kx] = (1+x)(1+x)^k - (1+x) - kx(1+x)
= (1+x)(1+x)^k - 1 - x - kx - kx^2
This line is almost exactly the same as the second line of my proof stage, expect it has an extra [-kx^2]. To compensate for this, we add kx^2 to the assumption multiplied by (1+x).
i.e. (1+x)<i>[(1+x)^k - 1 - kx]</i> <b>+ kx^2</b>
Assumption is in italics, and the "compensation" is in bold.
This way is also correct because the assumption is ≥ 0, (1+x) ≥ 0 and kx^2 ≥ 0, as it is a squared number.
I hope that makes sense. If not, just say so and I'll try to rephrase it =]