I am reading Khinchin's Continued Fractions page 10.
$\lbrack a_1;a_2,a_3\ldots\rbrack$ is a continued fraction and $q_k$ is given by $q_k=a_kq_{k-1}+q_{k-2}$. Suppose $\sum_{n=1}^{\infty}a_n$ converges so that there is a $k_0$ for which $k\ge k_0$ implies $a_k<1$.
Khinchin says "for $k\ge k_0$ we have $$q_k<\frac{q_l}{1-a_k} \, (*)$$
where $l<k$."
Problem: Since $a_k\to0, $ the right hand of $(*)$ side tends to $q_l$, which is fixed. However the left side tends to infinity. Since $q_k$ is a growing sequence, this cannot happen.
This looks like a bit of a looseness, perhaps in translation. The quoted inequality is not meant to be universally quantified in $l$. If you read the two paragraphs leading up to it, they show that either $q_k < q_{k-1} / (1-a_k)$ or $q_k < q_{k-2} / (1-a_k)$.
Thus the author simply means to say that $(*)$ holds for some $l < k$. If you read the inequality that follows you'll see it is applying this inductively to connect $q_k$ further back to a term earlier than $q_{k_0}$. This repeated application wouldn't be necessary if $(*)$ were true for all $l<k$.