THREAD: I was intrigued by the new "radicalization pipeline paper". Huge in scope, ambitious, I enjoyed it. I'm surprised no one has pointed out the biggest flaw though. Thought I'd unpack them from a Youtuber perspective.
Full paper 
https://arxiv.org/abs/1908.08313
-
Show this thread
-
The biggest shock was that the authors (with impersonal limitations) said you wouldn't find the alt-right from recommended videos. This reflects the important work Youtube has been doing in the past years since this problem had been noted, yet this was totally glossed over.pic.twitter.com/DNCBUJ1OpZ
1 reply 0 retweets 2 likesShow this thread -
Instead, the authors propose recommended channels from the sidebar of channel pages as a plausible mechanism of "radicalization". But if you know Youtube traffic like I do, you know NO ONE USES RECOMMENDED CHANNELS. I checked my own analytics. ~0.1% of my views come from that.pic.twitter.com/ZeL7fFQ11e
1 reply 0 retweets 4 likesShow this thread -
So this "pipeline", should it exist, seems to have been patched out by YT engineers, since really most traffic is driven by recommended videos, and literally a fraction of a fraction is driven by rec channels (the authors mechanism). Why does no one care about this fact!?
2 replies 0 retweets 2 likesShow this thread -
I believe YT has been getting a bad rap lately in the press (some deserved, some not), but we should also celebrate their wins. So that's why when the authors find that recommended videos do NOT lead to alt-right videos, that should be taken note of. Instead, the next sentence..pic.twitter.com/rKXBc1QGEq
1 reply 0 retweets 3 likesShow this thread -
...their "pathway" is the least likely way to find content on Youtube. I have to assume think this just reflects an unfamiliarity with how Youtube works, otherwise it feels like a non-sequiter. (I welcome the authors comments here:
@manoelribeiro ) Here's what we can learn..1 reply 0 retweets 3 likesShow this thread -
The authors demonstrate plausibility for *some* kind of mechanism IN THE PAST but fail to provide a plausible current mechanism for radicalization. Authors should entertain that Youtube has majorly fixed the problem, in terms of an impersonal pipeline feed of radical videos.
1 reply 0 retweets 4 likesShow this thread -
I know that upsets some people's narrative that
@SusanWojcicki and@TeamYouTube are ruining the world, but I think that given the data, this paper demonstrates the good work Youtube has done in addressing these problems. LMK what you think!1 reply 0 retweets 3 likesShow this thread
I'd also love to know what @JeffreyASachs thinks, given that I enjoyed some of his preliminary commentary on this paper.
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.