Live data from Hacker News

YouTube’s ‘dislike’ barely works according to new study on recommendations

engadget.com

421–430 of 437 posts

Re: YouTube’s ‘dislike’ barely works according to new study on recommendations

#421

Earlier quoted context omitted.

It wasn't that this player was disrespectful/racist/etc., it's that they were so much better that playing against this player was almost a guaranteed loss.

That sounds like a valid reason to avoid playing against that player.

Not in ranked games. I'd be the greatest chess player in history if I all I needed to do was block the grand masters.

Re: YouTube’s ‘dislike’ barely works according to new study on recommendations

#422
post #421

Earlier quoted context omitted.

That sounds like a valid reason to avoid playing against that player.

Not in ranked games. I'd be the greatest chess player in history if I all I needed to do was block the grand masters.

True, I suppose it depends on why you're playing the game in the first place. I was thinking of someone just playing them to have fun.

Re: YouTube’s ‘dislike’ barely works according to new study on recommendations

#423
post #96

Honestly, I feel like YouTube recommendation is the _only_ good thing YouTube has lol. I often see a ton of complaints about the recommendations but I've honestly never had an issue. I know that yes, you click on one video and now all the sudden you get 100s of that type... but it sorta works for me since it seems like that "effect" wears off just as fast as it starts. For example, I have a friend that's into cars. I…

Most people don’t understand here that folks at YouTube have probably ran thousands of variations of their algorithms and their current system is likely the best yet for keeping highest engagement on average . The issue might be that majority of the population might love recency bias. Folks here are not representative of general pop but they get overrun by majority preferences.

That's valid, but a bit depressing. It means that YouTube will never be able to accommodate me. Hmmm... maybe that's the realization I needed in order to finally give YouTube the boot. Thank you!

Re: YouTube’s ‘dislike’ barely works according to new study on recommendations

#424

Earlier quoted context omitted.

Why is everyone pretending that the maximal amount of ad revenue comes doesn’t come from showing you what you want to see? Individual videos don’t pay to be in the recommended feed. There’s no opportunity to see an ad except when you watch a video. YouTube has a perfectly balanced incentive to show you more content you’ll watch. Where is the exclusive choice between (what the user wants to watch) and (what makes more…

Technically correctly, although there is a subtle nuance here. What you want to see is not necessarily the same as what you will end up watching. Just because you dislike a video doesn't mean you wouldn't watch it. Or put another way, you can tell YouTube what you want (to be perceived socially as wanting) to see, but the algorithm knows what you actually want to see.

> Or put another way, you can tell YouTube what you want (to be perceived socially as wanting) to see, but the algorithm knows what you actually want to see.

Except the algorithm doesn't know what I actually want to see. Nobody but YouTube knows what I tell YouTube I like or dislike, so there's not even a theoretical pressure to lie about by wants in order to present an appearance to others.

Re: YouTube’s ‘dislike’ barely works according to new study on recommendations

#425
post #148

I'd be curious to know how many additional "wanted" recommendations the system is able to serve by not treating a dislike as a strong negative signal. It seems like there's a tradeoff here, where YouTube figures that if it interpreted a dislike with a heavy hand, it would also knock a bunch of videos you do want (or at least they think you want). Presumably from YouTube's perspective, it's more important that your re…

Except that when the recommendations are 90% garbage, I'm not going to sift through them to find the occasional gem. That's just burning time for little reward.

Re: YouTube’s ‘dislike’ barely works according to new study on recommendations

#426

In practice it doesn't work in the short term but it should work in the long term because Google processes so much data with billions of data points so their algorithms need some time to catch up. Sad but true. The same situation is with search ranks; they update their ranking index every few months or so according to billions of data points they gather with the final word of ranking algorithms. Basically every big i…

> In practice it doesn't work in the short term but it should work in the long term

In practice, it doesn't work in the long term either. I've been using YouTube regularly for many years.

Re: YouTube’s ‘dislike’ barely works according to new study on recommendations

#427

I wish I could dislike entire genres of things, specifically reaction videos. Laziest “content” ever and they completely takeover results anytime I search for music.

For me, it's those damned "shorts". I really wish I could just make those completely invisible.

Re: YouTube’s ‘dislike’ barely works according to new study on recommendations

#428
post #336

Earlier quoted context omitted.

It's probably companies advertising their products getting salty their new movie or something got 90% downvotes and YT is all too eager to please its ad overlords.

It started when a Joe Biden speech video got disliked into oblivion. We're talking 2 million+ dislikes.

Why not do things transparently and do weighting by default weight = upvotes - downvote + us cents paid by author, all 3 visible.

Re: YouTube’s ‘dislike’ barely works according to new study on recommendations

#429

Earlier quoted context omitted.

YouTube doesn't show you ads if you buy Premium.

You need the YouTube Pro Max+ Platinum bundle to get functional search.

Yeah, search doesn't work at all. It does seem like it pulls up specific channels if you put "ch" in the search text, so I tend to go from there.

Re: YouTube’s ‘dislike’ barely works according to new study on recommendations

#430

Earlier quoted context omitted.

I used to have the same experience of YouTube recommending just more of whatever I just watched, and at most recommendations would stick around for 6 weeks before it forgot about a subject. That was until earlier this year when I watched some videos from MrBallen (camp fire stories based on real events), and ever since then I have gotten my recommendations filled with crazy conspiracy stuff. It stopped recommending h…

If you want to fix it (and I agree it shouldn't be this painful, but you deal with what you got), you can go back in your history and remove the videos from MrBallen from your watch history. This should fix it for good, almost immediately.

Thanks
Post reply on HN