> then you can account for that and not have the algorithm register them
It’s an unsolvable problem which is why they disabled it entirely. If you “account” for “bad” input the only consequence is that those responsible for that bad input figure out how to get it classified as good input.
If it's an unsolvable problem for downvotes, it's an unsolvable problem for upvotes, too. The reason they took away downvotes is because they started to partner with the networks, to artificially boost their posts, and to deemphasize and demonetize their traditional amateur comment (inspiring a shooting.)
The mainstream content gets ruthlessly downvoted because polish doesn't equal quality, and the networks (wisely) don't want their stuff distributed by a platform that allows users to mark it as bad. So Youtube took away the ability to mark content as bad. It's no more complicated than it looks.
I am sure the removal had more to do with certain ideological views/content being more likely to be downvoted. The decision came soon after the critically panned Susan Wojcicki YouTube CEO 2021 Free Expression Awards. This probably has the record for the worst ratio of any video in the site's history.
What? Suddenly we are in a ficitious world where there are droves of trolls disliking videos, and somehow they are sentient as to YouTube's recommendation algorithm, and they desperately want the video to disappear from everyone's recommendations, so these trolls come up with new and inventive ways of DISLIKING videos?
I don't think I've ever read more made-up scenarios than on this website.
Google didn't think it'd be worth it to pursue a solution that would solve this, and probably nobody internally wanted ownership of it. The second best solution is the one that doesn't cost much and solve the problem. Youtube being the only game in town, where are people gonna go anyways? Daily Motion? Post implementation KPI probably showed that traffic hasn't budged, and the problem has been solved. As far as Google is concerned, it was a rational decision.
You don't even need to be famous in order to get trolls who are hating you so much that they make it their job to ruin your life. When you get above a certain size you will have organized troll armies coordinating attacks on private Discord servers.
right, I don't think the comment was disregarding bregading. That a multi-billion dollar company's algorithm is too dumb to deal with it is what astounds.
You basically can't allow people to give negative feedback for a thing and have that feedback mean anything (ie affect recommendations for anyone but you or show it to other users) without insincere feedback being used to hurt the reviewee.
There are ways to counter this, the easiest is to not show negative reviews but count them positively, a dislike actually boosts them just like a positive review would. Not really recommended due to promoting rage bait but brigading would stop working.
You also can't allow people to give positive feedback without insincere feedback being used to artificially promote things. If you're going to have a feedback system, you need some manner of checks and balances on it to limit insincere feedback.
> There are ways to counter this, the easiest is to not show negative reviews but count them positively, a dislike actually boosts them just like a positive review would,
I think it would be better to just merely count the number of upvotes and complete, or almost complete views, for boosting purposes.
The downvotes should be for tailoring feeds, whether personal, or the aggregate feeds of people with similar interests and like/dislike votes.
Any specific criticism can be saved for the comments.
It’s an unsolvable problem which is why they disabled it entirely. If you “account” for “bad” input the only consequence is that those responsible for that bad input figure out how to get it classified as good input.