Only commenters with a certain P-score are allowed to post. The AI also determines if the comment is negative or positive, and will auto-block it if it's negative.
You seem to be knowledgeable about this. All I know is they have some weird secret rules when it comes to comments on videos and in live chats, and they are targeted towards certain users. For example I cannot post comments with innocuous words like "death" or "bullshit". The comment will be shadowblocked I guess - it appears to me but nobody else, and is gone when I refresh.
However if I post the exact same comment under a different account (or ask someone else to test) it goes through. That person, on the other hand, get blocked when using different words.
There are companies working for Big Tech who hire contractors under NDAs to do work to "train" the AI based on certain word phrases, rebuttals, responses, and sentiments considered "emotionally upsetting".
This is across the board for both speech and text.
You're right that certain words/phrases will net you a shadowblock on one account but it won't affect the other, unless they say something disruptive, or what they label as "emotionally upsetting."
The way it works is identical to the way they rate actual YouTube channels. Have you heard about P-scores? I think you can search it up on Yandex or DuckDuckGo and there are some articles/videos from non-Leftists explaining how the system works.
When working under contract they don't tell you about the P-scores, but the way the system works it's obvious how it relates to said P-scores.
The short gist of it is that when you make comments that are labeled as "positive", it adds to your score. It doesn't mean you can just say anything, but it does grant you more leniency in having your comment appear on videos or about topics that relate to Liberal Progressive subject matter.
If score is too low then your comment (or rather, your account) is completely shadowblocked from posting on certain channels covering certain subject matter.
If you link to a certain site, if you mention a word, or word phrase that they have trained the AI to consider to be "derogatory" or "offensive", then your comment will automatically be removed.
If your profile identifies you as a certain race and you attempt to make a negative comment on a video made by or for another race (e.g., black or Hispanic) then your comment will automatically be removed or shadowblocked along with your account. If it's a first-time offense you may just have the comment blocked and receive a negative P-score.
The reality is that you have no idea how many people are actually being blocked, because the AI is censoring a lot of comments in real-time once certain banned words, phrases, sentiments, talking points, links, references, or arguments are made.
You can sometimes get around this by using brackets [] or parenthesis () and putting them around certain words to get through the filter, such as: [f]emin[i]st [a]gen[d]a or (g)a[y] [a]ge[n]d[a].
But it makes for very cumbersome forms of communication. And eventually the AI will be trained to weed out these kind of comments.
Sadly this is the future of how Big Tech will control conversations and people need to be very aware about how they disseminate information across Big Tech controlled platforms.
Shadowblocking.
It's like that for a lot of videos.
Only commenters with a certain P-score are allowed to post. The AI also determines if the comment is negative or positive, and will auto-block it if it's negative.
You seem to be knowledgeable about this. All I know is they have some weird secret rules when it comes to comments on videos and in live chats, and they are targeted towards certain users. For example I cannot post comments with innocuous words like "death" or "bullshit". The comment will be shadowblocked I guess - it appears to me but nobody else, and is gone when I refresh.
However if I post the exact same comment under a different account (or ask someone else to test) it goes through. That person, on the other hand, get blocked when using different words.
YouTube is fucked up.
It's all deliberate.
There are companies working for Big Tech who hire contractors under NDAs to do work to "train" the AI based on certain word phrases, rebuttals, responses, and sentiments considered "emotionally upsetting".
This is across the board for both speech and text.
You're right that certain words/phrases will net you a shadowblock on one account but it won't affect the other, unless they say something disruptive, or what they label as "emotionally upsetting."
The way it works is identical to the way they rate actual YouTube channels. Have you heard about P-scores? I think you can search it up on Yandex or DuckDuckGo and there are some articles/videos from non-Leftists explaining how the system works.
When working under contract they don't tell you about the P-scores, but the way the system works it's obvious how it relates to said P-scores.
The short gist of it is that when you make comments that are labeled as "positive", it adds to your score. It doesn't mean you can just say anything, but it does grant you more leniency in having your comment appear on videos or about topics that relate to Liberal Progressive subject matter.
If score is too low then your comment (or rather, your account) is completely shadowblocked from posting on certain channels covering certain subject matter.
If you link to a certain site, if you mention a word, or word phrase that they have trained the AI to consider to be "derogatory" or "offensive", then your comment will automatically be removed.
If your profile identifies you as a certain race and you attempt to make a negative comment on a video made by or for another race (e.g., black or Hispanic) then your comment will automatically be removed or shadowblocked along with your account. If it's a first-time offense you may just have the comment blocked and receive a negative P-score.
The reality is that you have no idea how many people are actually being blocked, because the AI is censoring a lot of comments in real-time once certain banned words, phrases, sentiments, talking points, links, references, or arguments are made.
You can sometimes get around this by using brackets [] or parenthesis () and putting them around certain words to get through the filter, such as: [f]emin[i]st [a]gen[d]a or (g)a[y] [a]ge[n]d[a].
But it makes for very cumbersome forms of communication. And eventually the AI will be trained to weed out these kind of comments.
Sadly this is the future of how Big Tech will control conversations and people need to be very aware about how they disseminate information across Big Tech controlled platforms.