Incredible proof of professional degeneracy from a certain type of academic.
This paper is likely the best example of someone with ultimate credibility on a topic being OBLIGED to post a paper. When an active X (formerly, and generally referred to as Twitter) user is asked to post a paper, this is what you get.
The paper is full of gems and is almost unintelligible. It is a must-read for anyone interested in the state of research (2018) in visual classification. I am by no means an expert, but I cracked up multiple times reading it.
Here is one quote to set the stage:
'Anyway, this paper won’t have become what it will in time be without all the work your past selves will have done also in the past but only a little bit further forward, not like all the way until now forward. And if you tweeted about it, I wouldn’t know. Just sayin.'
Arxiv link: https://arxiv.org/pdf/1804.02767
Follow the author on X: @pjreddie