After trying to understand the accepted answer, I have decided to rewrite the answer in my own words:
By the definition of limits, to show $\lim_{x \rightarrow 1}x^3=1$ we must show that for all $\epsilon > 0$ there exists $\delta > 0$ such that whenever $0 < |x-1|<\delta$ we have $|x^3-1| < \epsilon$.
Let $\epsilon > 0$, and let's look at the statement $|x^3-1| < \epsilon$. Notice that $|x^3-1|$ = $|x-1||x^2+x+1|$ and so the statement holds if $|x-1||x^2+x+1| < \epsilon$. At this point, we already have an $|x-1|$, and so we are close to choosing $\delta$ in terms of $\epsilon$, but we must get rid of $|x^2+x+1|$ first.
The statement holds if $|x-1|<\frac{\epsilon}{|x^2+x+1|}$. We would like to get only a constant on the right hand side, so is there anything we can say about $|x^2+x+1|$?
Notice that if we consider some number $r$, either $|x-1| \leq r$ or $|x-1| > r$. Let's arbitrarily choose $r=1$ (any choice should be fine here), and so either $|x-1| \leq 1$ or $|x-1| > 1$. We would need to show that our single choice of $\delta$ takes into account both possibilities.
Why is this useful? Let's look at the first possibility. We know that $|x-1| \leq 1$ holds if $-1 \leq x - 1 \leq 1$, which holds if $0 \leq x \leq 2$. But also, if $x$ lies in this range, we have $|x^2+x+1|\leq|2^2+2+1|=7$. So, $\delta=\frac{\epsilon}{7}$ would be a good choice so long as $|x-1| \leq 1$.
But, we need to find $\delta$ that works for the other values of $x$ also, since that was only one of the two cases for $x$!
However, we are free to choose the smallest $\delta$ we like. Can we make $\delta$ small enough where $|x-1|$ is never greater than $1$? That way we wouldn't even have to worry about the other case!
Notice that since $|x-1|<\delta$, to ensure $|x-1| \leq 1$ we just need to ensure $\delta < 1$.
Choose $\delta=\text{min}\{\frac{\epsilon}{7}, 1\}$. $\space\space\space\square$
This can probably become a common strategy: Make $\delta$ small enough so that the absolute value is less than an arbitrary real number $r$. Once that's the case, you can find an upper bound for $x$, and this allows one to get rid of all the extra $x$ terms.