I think I may have discovered a bug with WolframAlpha.
So I was trying to determine all $x$ such that
$$\sum_{i=0}^{5}{x^{-i}} < \frac{13}{12}.$$
WolframAlpha spit out $x > 13$ (see this link).
However, when I double-checked,
$$\sum_{i=0}^{5}{{13}^{-i}} = \frac{402234}{371293} \approx 1.0833331\ldots.$$
Of course, we have
$$\frac{13}{12} = 1.08\bar{3}.$$
This means that $x = 13$ is a solution to the inequality
$$\sum_{i=0}^{5}{x^{-i}} < \frac{13}{12}.$$
Here are some more examples:
Does this mean that we can't trust WolframAlpha?
Yes. A simple equivalent is trying to solve:
$$13x^5<x^6+12$$
So you are seeking solution to $f(x)=x^6-13x^5+12=0$. $x=13$ is a gross estimate, but it is not far off. If we take $x=13+\epsilon$ then $(13+\epsilon)^6-13(13+\epsilon)^5 + 12 = 0$ which, if we ignore exponents for $\epsilon$ greater than $1$ yields the equation:
$$13^6 + 6\epsilon \cdot 13^5 - 13(13^5+5\epsilon\cdot 13^4) + 12 =12 + \epsilon \cdot 13^5$$
So $$\epsilon\approx \frac{-12}{13^5}\approx -0.00003231948892$$
And even that is not perfect: $f(12.99996171721108)\approx -2.21$.
A more accurate answer is: $$x> 12.999967680123.$$
Then $$f(12.999967680123)\approx 0.0000050783$$
I don't know what WA used for rounding. It only appears to use 6 digits of accuracy elsewhere, which would result in $x>13$.
If I ask WA to:
It returns $13.0000$.
When I ask WA to:
it returns $-0.0000323199$.