Ok, it isn't a perfect proof. However, what are the odds (considering how many people are writing compared to any specific reader) that increasing reading speed will enable writers to research, proofread, copy edit, and revise faster?

Nearly certain, but the point is that even outright assuming certainty doesn't lead to the conclusion. Odds aren't the issue, relative degree of improvement is. For example if writing_time is equal to some multiple of reading_time plus a 'creative effort' constant and reading speed improves sufficiently then someone will be able to read all of the books despite a drastically and ongoing increase in books written.