Read on Twitter
Thread on Youtube censorship.
Grabbed ex-YT friend (anonymized) to discuss today. They disclaimed that they do not work directly on this, but their hypothesis makes more sense to me than Conspiracy Twitter (Original convo QTed, where I make some guesses that I still stand by) 1/ https://twitter.com/eigenrobot/status/1265201959096606720
Grabbed ex-YT friend (anonymized) to discuss today. They disclaimed that they do not work directly on this, but their hypothesis makes more sense to me than Conspiracy Twitter (Original convo QTed, where I make some guesses that I still stand by) 1/ https://twitter.com/eigenrobot/status/1265201959096606720
We (friend + me) currently model the anti-spam anti-bot stuff at Twitter as an interlocking system, which lots of different levers that inform each other. We do not think there's just some monolithic dude that just sits and censors words, since that doesn't scale. 2/
My original hypothesis in original thread is that this is automated. My friend agrees that there are probably algos all over the place learning different things like what's noisy, what pattern-matches to bots, etc., so this is probably literally an accident from YT's POV. 3/
This doesn't mean there's no agency or that PRC didn't influence it. Some examples:
a) maybe a lot of people (possibly PRC-coordinated) simultaneously flagged comments as abuse that just had these words in it. If this is the case, these anti-CCP words may be learned 4/
a) maybe a lot of people (possibly PRC-coordinated) simultaneously flagged comments as abuse that just had these words in it. If this is the case, these anti-CCP words may be learned 4/
b) maybe there's an internal bot that "learns" spam accounts, and also learns that spam accounts spam these types of anti-CCP stuff, and blocks that through *another* bot that censors words said by spam accounts that are considered malicious 5/
c) I *totally* think it's possible that some people have manual override powers, and through social engineering they got some person with plausible deniability to hand-censor a word. However, this leaves trails and I think it's less likely 6/
Finally, what I do *NOT* think happened is that the CCP straight up asked YT and they kowtowed. YT demonitizes a lot of pro-CCP content, does *not* have market share in China and doesn't care, and has shown no real political reason to bow to China in this way. No leverage. 7/
Summary: I think this is literally *automated*, possibly CCP-influenced by actions like mass-comment flagging, but no way a conspiracy. Criticizing YT is definitely barking up the wrong tree and feeds into the "tech bad" mind-killer. Real sycophants are elsewhere. 8/8 END
Addendum 1: sorry for not being clear in opening tweet.
Chronology:
1) last night we ended up discussing this, as a tangent in QT-ed thread
2) today I had a private convo with friend (not the thread QTed in OP)
3) I synthesized (2) in this thread, providing (1) as context.
Chronology:
1) last night we ended up discussing this, as a tangent in QT-ed thread
2) today I had a private convo with friend (not the thread QTed in OP)
3) I synthesized (2) in this thread, providing (1) as context.
