It's easy to prove that if $I$, $J$ are two-sided ideals and $R/I\cong R/J$ as modules over $R$, then $I=J$. What about left ideals? Is there a simple counterexample?
I believe I've found an answer, but since answering own questions is encouraged, I thought I might post it here. Other examples are obviously welcome.
When I try to find (counter)examples, I don't consider random examples first, but rather try to simplify the general case so that, in the end, examples pop out automatically without any effort. This also works here:
By the universal properties of $R$ and quotient modules, a homomorphism of left $R$-modules $R/I \to R/J$ corresponds to some element $[x] \in R/J$ such that $Ix \subseteq J$.
Hence, an isomorphism $R/I \cong R/J$ is given by two elements $x,y \in R$ such that $Ix \subseteq J$ and $Jy \subseteq I$ and $xy \equiv 1 \bmod I$ and $yx \equiv 1 \bmod J$.
For example, it could happen that $y=x^{-1}$. But then the conditions reduce to $Ix=J$.
Thus, any left ideal $I$ of a ring $R$ with a unit $x$ such that $Ix \neq I$ provides a counterexample.
As mentioned by Marcin Łoś and Andreas Caranti, we may look at $R=\mathrm{End}(V)$ for some vector space $V$ and $I=\{f : f(v)=0\}$ for some vector $v$. If $x \in R^*=\mathrm{GL}(V)$ is chosen such that $v,x(v)$ are linearly independent, then $Ix \neq I$.