Screenshots to prove that it was < 1000 at the time this post is made:
If you look at their activity, you see that the badges were also awarded in quick succession (with the order depending on which script ran first).
This means that the views would've had to come in rapidly in large volumes that it managed to beat the scripts and bucked the natural pattern of Announcer being awarded first, Booster a few hours/days later and finally the Publicist badge.
This is not normal viewing behaviour where someone chances upon a link and then maybe clicks it and moreover, the question isn't that interesting/catchy to have such a large volume of views. If the devs look at the kurtosis of the data points (time viewed), I'll bet they'll find a very high kurtosis for this case.
My guess is that the strict view counter™ also takes into account the velocity of views, which explains the low views displayed (closer to the actual number), whereas the link tracking script merely looks at unique IPs, making it easy to farm badges if you have the means and knowhow to write a script to do that.