The size of the men of the pen in the Ayyubid dynasty (and premodern islamicate polities)

Between 5 and 8 million people are estimated to have lived under the Ayyubid dynasty at its height. Just a small number of people were identified as "knowledge makers" from this vast population. Only a handful did belong to the literati. Who can be defined as the literati? I have a slightly higher standard in mind than merely being able to read, which is impossible to quantify for so many centuries ago. But it is possible to quantify the higher standard. Some examples to get a feel for the group I have in mind: authors, physicians who studied the "ancient sciences" (Greek heritage sciences), government scribes, the literati with a position at some institution etc. Let's use the most liberal estimate, which would be... a few thousand. 

When people read texts about the premodern intellectual history of science and philosophy, they are led to believe that these dynasties had philosophers and scientists who had fallen from trees in large numbers at some point. That was not at all the case in fact. 

Nevertheless, their historical influence is fascinating. How can we contextualize the estimation for the contemporary era? The cognitive ability of these well-known individuals can be roughly estimated. In today's developed societies, a few thousand literate people are not very significant at all. As an illustration: Consider the big-cap firms that are present in many developed nations. Tens of thousands of people work for many of them. If there are only a few thousand "high achievers" or even fewer, it is impossible to maintain properly functioning modern societies with their growing complexity. As a result, in the modern world, it would be challenging for such a society to become notable in worldly criteria. 

The bigger argument I'm trying to convey contradicts two lines of perspective. One of them consists of individuals questioning why premodern Islamic polities performed so well intellectually and how they differed drastically from modern-day awful performance of Muslim countries. Another line of reasoning may incorrectly assume that Islamicate polities had a large number of people who could count to more than 10. Both assumptions are comically incorrect, given how rare these individuals were, even as a whole throughout all of islamicate history, let alone for specific dynasties throughout Islamic history.