Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

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

engadget.com

Post reply on HN
Print view View original

391–400 of 437 posts

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

#391

Post by JohnFen »

hombre_fatal wrote: post #102
JohnFen wrote: post #49

Earlier quoted context omitted.

> With that in mind, I think Youtube recommendations are quite good. Having to nuke a valuable reference like your history in order make the recommendations approach something useful is the opposite of "quite good".

Well, I'm one to appreciate what we have available to us especially compared to the alternatives and temper my appreciation accordingly. I don't know if I use a single service that gives me particularly compelling control over recommendations. Though I also don't find myself deep-diving my history so often that clearing it every few months affects my life, but I suppose I can see how that could be a setback.

My main point is that if the recommendations are improved by giving YouTube _less_ information to work with, that's a solid sign that the recommendation system is borked.

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

#392

Post by vineyardmike »

wodenokoto wrote: post #273
vineyardmike wrote: post #129

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…

> Why is everyone pretending that the maximal amount of ad revenue comes doesn’t come from showing you what you want to see? There's a very philosophical debate hidden between the words in that sentence. But since Youtube sees dislike as a signal to boost a video, it spreads content-views to things that people don't like, but feel compelled to interact with in (what I assume is) the "Not now, I'm busy - someone is be…

> But since Youtube sees dislike as a signal to boost a video, it spreads content-views to things that people don't like, but feel compelled to interact with in (what I assume is) the "Not now, I'm busy - someone is being wrong on the internet" kind of way.

I just don’t see YT having the same feedback loops as other places. You can stay on without being directly typing in comments like on fb or twitter.

I personally never see rage-inducing content on YouTube (nor other sites), nor do i see it on my partners, friends, parents accounts. I’m sure it’s there, but I think that people who get it are training their recommendations to show it, not being forced it. I get way more rageful content on TikTok ironically, which has a way weaker “dislike” signal and a stronger propensity to show you new things.

To the point, I understand the incentive someone like fb has to show content people don’t like but still engage with, but I’m not sure that the incentive exists in YouTube. Furthermore I think we discredit companies (who are made up of sentient and capable people) to recognize short term vs long term value and the importance of reputation.

I reject the notion that there’s someone in San Bruno pitching that they use the dislike button to show more of that video instead of a negative quality signal. No one I know in Silicon Valley who works in tech would make such a suggestion, and they’re no different than the YouTube team on a fundamental level.

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

#393

Post by vineyardmike »

Ferret7446 wrote: post #205
vineyardmike wrote: post #129

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.

Well yea, but that’s the wrong conclusion.

By definition, clicking on a a dislike button means you watched at least part of the video. It’s a signal that “I started watching this and it wasn’t a good match”.

That should send the message that it’s appealing from the title/thumbnail but a bad video. I don’t think the “socially be seen as wanting” plays too much here… because starting to watch already. You already sent a different signal.

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

#394

Post by mrguyorama »

xani_ wrote: post #343
johnfn wrote: post #67

Earlier quoted context omitted.

> I watch it daily, I have never listened to any hip hop, I don't watch sports nor trending topics, yet every single day it suggests football videos, MMA, the latest hiphop hit, and celebrity crap (imagine the Kardashians) among my other interests. I don't have this experience at all. YouTube recommends me pretty much exactly what I've been recently watching. I watch a lot of long form interviews and tech videos, and…

I got a lot of recommendations for stuff I ALREADY WATCHED, it even shows the red bar showing it KNOWS I watched it, still recommends

Youtube does that because people like me compulsively re-watch content. Youtube probably needs a few user-visible knobs like "I like to rewatch stuff vs I hate rewatching stuff"

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

#395

Post by mrguyorama »

sph wrote: post #284
fallingknife wrote: post #42

Earlier quoted context omitted.

It's insanely biased towards popularity. I watched this series of like 40 videos on extremely large numbers. And not one time did I ever see one in my recommendations. You would think when you've watched 1-20 of something the algorithm would be smart enough to suggest 21 but no.

What's funny is that the super smart AI behind the rec engine in YouTube sometimes is so dumb it can't figure out that video #3 in a series with consistent titles should be followed by video #4. Sometimes it just gets it wrong and skips one, other times it just recommends some other totally unrelated video even if you've been watching the same numbered series forever. And it's got a short enough memory span that one…

I wonder how many of the "algorithm's" problems are caused by a huge number of literal toddlers that just hit buttons at random. How many parents just sit their kid at an iPad for some time to do something?

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

#396

Post by adolph »

Bakary wrote: post #334
adolph wrote: post #231

Seems like someone has rediscovered the: Placebo Button is a push-button or other control that appears to have functionality but has no physical effect when pressed. Such buttons can appear to work, by lighting up or otherwise reacting, which rewards the user by giving them an illusion of control. They are commonly placed in situations where it would have once been useful to have such a button but the system now oper…

The ballot box as well. Only half joking here.

The flag link for election misinformation....joking

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

#397

Post by TheMode »

daveoc64 wrote: post #131
mikrl wrote: post #80

I didn’t like when they removed the dislike counts, and I still don’t like it. When I look for videos of folks working on Subarus, I don’t want to have to manually scrub through the video to gauge the quality. Before I’d just look at the ratio and if I was feeling generous I’d preview the frames to see if they had what I wanted. Now it takes me at least 2 mins longer to get to the same judgement which begs the questi…

The problem with the dislike button seems to be that nobody agrees on what it should be used for. In previous HN discussions, it's become apparent that some people use it to say "I am not interested in this video". That is of zero use to other users. The video isn't necessarily bad, it's just on a topic you don't care about. If you are instead using it to mean "this video is factually wrong" or "this video is obnoxio…

Fact is that it already worked well, no point in trying to find unnecessary confusions

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

#398

Post by kavenkanum »

Garvi wrote: post #326

About 6 months ago I ran my own little unscientific experiment with Youtube, after I noticed them promoting a certain channel that I've never watched. Nothing political or anything like that. Just a run of the mill youtuber, currently very popular with the kids. The first week or two I was pushing the "not interested" and "don't recommend channel" every time one of the videos would show up in my feed (mind you I've n…

Yeah, flagging stuff or just clicking dislikes sometimes does not work at all. I feel this is due to popularity being more important than these dislike acts. Additionally Youtube algorithm also treats user `dislikes` like an engagement, so it may work against your common sense.

What I did find very useful some time ago was to actually `Block` whole YouTube channels (you do it from Channel's Info page by clicking on the Flag button). This is an ultimate way of getting rid of some content on your feed. I blocked pretty much all these 5-Minute Crafts, Blossoms, Bright Side and other crap-generating trolls. Best decision ever!

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

#399

Post by whywhywhywhy »

vonwoodson wrote: post #307

They left out the block channel button. Very effective. Seriously, as an avid YouTube consumer (and not a creator) training my algorithm is such a huge part of what my entertainment life is. Definitely have used private mode, on a VPN, with a different browser to watch videos I didn’t want polluting my algorithm. Now that my kid knows how to use YouTube and I let him use my account, my years of work are totally shot……

The “Block channel” does nothing as a viewer. It just means that actual YouTuber can’t comment on your videos.

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

#400

Post by vineyardmike »

ax0ar wrote: post #178
vineyardmike wrote: post #125

Earlier quoted context omitted.

How does that at all relate to YouTube videos though? YouTube wants you to see ads. Ads are put in front of YouTube videos. Therefore youtube wants you to watch videos. YouTube, logically, should want you to watch more videos, for longer . If they serve you bad videos, you stop. Therefore, Googles incentives are to tailor good recommendations of content for customers. QED. I think all this “you’re not the customer yo…

They are sort of recommending 100 videos to everyone. In my circle, a lot of people I know have very different interests than me. Yet, whenever I recommend a video to a friend, they often say it was already in their recommendations. They might be pushing videos or topics that advertisers are more willing to pay for.

> a lot of people I know have very different interests than me. Yet, whenever I recommend a video to a friend, they often say it was already in their recommendations.

Is it possible that Youtube knows your friend's video interests just like you do?

> They might be pushing videos or topics that advertisers are more willing to pay for.

This, however, does seem very possible.

Post reply on HN