I want to train it to give me communist content, news, culture discussion from the leftist perspective, ect…basically this site. Never used it before so I’m asking before I do anything on it so I don’t mess it up with the initial searches or anything.
Search by hashtags, and go like prolewiki videos to teach the algorithm 🥰
If you don't watch a video more than a few seconds (swipe down), the algorithm takes that as a signal you don't want to see this
If you like a video, the algorithm takes it as a sign you like the video and want to see more.
In my experience, it seems if you upload videos with a certain hashtag, you'll start seeing more of that hashtag.
follow creators in the genre you like too
Nowadays my fyp is mostly communist stuff but there's still the odd fascist shit that seeps through, because the algorithm is not quite there yet in differentiating.
Someone else said the fascist stuff is there to try and rage bait you. Maybe it is trained and that’s why because it knows you’ll hate it?
Honestly anything is possible with social media, the only way we know how algorithms work is by trying to reverse engineer them. Companies don't really publish how they work as it's a trade secret.
Exactly. All we know for sure is that it is driven by 'engagement' and that basically means all of the old tabloid tricks are also on the table alongside the newer digital engagement models that were pioneered by sites like Gawker, and we know based on interviews that Gawker used outrage as one of its primary engagement models.
TikTok loves a rabbithole but it will constantly push other stuff just to keep testing if you want to engage with the other side of the spectrum.
That’s true, glad I made the thread though. Learning a lot of stuff so far. Going to keep in mind that the exact mechanics are secret though, it’s a good point.
Probably just the statistical pattern of engagement. People who like communist content are probably likely to troll/argue with/debunk or otherwise engage with fascist posts by commenting on them, and the algorithm won't be able to tell the sentiment of the comment without costly and mostly useless sentiment analysis. These type of algorithm generally do not "know" anything about the meanings of specific tags, only the statistical relationships that emerge in how users who like them interact with other posts. Which is funny because it means that it can do rage bate simply because it works without any human ever intending it to (but of course, no human stopping it either).