Apologies if these are not typical definitions - I'm new to set theory, but one thing has been really confusing me for a while.
Define a real number $x$ to be definable if it can be uniquely specified in ZFC. In other words, there is some finite string of symbols in ZFC which is satisfied by $x$ and only $x$. Since there are only a countable number of finite strings in ZFC, cardinality of the definable real numbers is countable. Suppose that $(x_1, x_2, x_3 \dots)$ is the list of all definable real numbers ordered by the shortest equation that defines them.
Now, apply Cantor's diagonalization argument: let $d_i$ be $1$ if the $i^{th}$ digit of $x_i$ after the decimal point is $0$, and $1$ otherwise, and let $x = 0.d_1d_2d_3\dots$. Since for every $i$, $x$ and $x_i$ differ in at least one digit, $x$ is not a definable real number. On the other hand, I've just given an explicit, uniquely defining description of $x$. So $x$ should be definable.
What's going on here?