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

41–50 of 437 posts

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

#41

Post by svnpenn »

TheAceOfHearts wrote: post #32

The trick to improving YouTube recommendations is to regularly prune uninteresting videos and channels. Eventually this should start surfacing videos you're more likely to find interesting. I've actually hit the end of my recommendation stream on the mobile app after marking dozens of videos with Not Interested. Once this happens it starts recommending videos you've already watched or older videos from creators you'v…

> I'm never interested in rewatching stuff I've already seen.

thats a weird position. The whole point of YouTube (used to be anyway) was long form content. YouTube is not new anymore. You really have no interest in watching some good 10 minute video you saw 9 years ago?

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

#42

Post by fallingknife »

sph wrote: post #17
HardwareLust wrote: post #5

YouTube doesn't seem to understand my preferences, at all. It's far, far too biased toward what you recently watched. It basically doesn't work and certainly doesn't "learn" anything.

YouTube absolutely IGNORES your preferences. 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 click "don't recommend this channel", or "I don't like this video " every single day, I have an extension to make this conveni…

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.

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

#43

Post by jdmdmdmdmd »

kache_ wrote: post #33
makosdv wrote: post #29

I'm not surprised by this. I gave up on the YouTube Home page, etc. years ago and I only look at the Subscriptions page. I've just found new channels to subscribe to through word of mouth over the years and I have more than enough content to choose from.

Which is actually quite a good way to find youtube channels! I find them through twitter, 4ch, and friends.

>twitter, 4ch, and friends

I was under the impression that these three are mutually exclusive to each other

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

#44

Post by SllX »

TheAceOfHearts wrote: post #32

The trick to improving YouTube recommendations is to regularly prune uninteresting videos and channels. Eventually this should start surfacing videos you're more likely to find interesting. I've actually hit the end of my recommendation stream on the mobile app after marking dozens of videos with Not Interested. Once this happens it starts recommending videos you've already watched or older videos from creators you'v…

You’ve nailed most of the curation, but the way I get new recommendations is that I search out new stuff on YouTube. At this point if there’s any subject matter I’m interested in that’s not political or social issues, I’ll do a search on YouTube in addition to my other searches (Google, Brave Search, Reddit, etc.) to see what I can turn up. This often nets me additional recommendations down the line, often when there’s a new video on that subject posted when I’m not expecting it.

Low quality videos also get nuked with prejudice from my history.

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

#46

Post by proteal »

amatecha wrote: post #19

This reminds me of how games ( cough LeagueOfLegends cough ) match me up with a player that I literally just reported for harassment, griefing etc. in the previous match. Like, there weren't any other possible players to match me with out of the many thousands? I don't mind waiting a few more seconds to match up with someone else! I'm getting a lot of responses as if I expect the player to be totally blocked from mat…

Overwatch had a feature where you could avoid players, but a top sniper was avoided by enough top players that he could never find a match. The matchmaking system could never find him a match because almost all potential matches had at least one player on each team that had him avoided. The devs eventually reworked that feature so that he could play the game. Sometimes seemingly innocuous features have weird edge cases!

Even for an average player, adding an avoid player feature would ultimately make matchmaking take longer. It necessarily reduces the number of valid matchings of players, so the system will only be slower. My guess is that average queue time is an important metric that management does not want to increase. It’s an especially tough sell because if you really wanted fresh players in your game, you could always not queue for a few minutes and you will almost certainly get all-new teammates.

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

#47

Post by mrkramer »

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 internet service runs on algorithms. You name it: Ebay, Amazon, Google, Facebook, TikTok, Netflix, Spotify etc.

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

#49

Post by JohnFen »

hombre_fatal wrote: post #11
beart wrote: post #6

Earlier quoted context omitted.

I feel like every recommendation service works this way - YouTube, Amazon, Spotify, etc. Whatever the last thing you looked at becomes your entire life's mission according to these systems. When I think about all the buzz words around machine learning, ai, and so on, it's completely incompatible with the reality I see. What exactly are all these engineering teams doing just to end up with this result? Amazon might be…

Youtube does have a great escape latch though: go to the History tab and delete your viewing history. Your recommendations start from scratch and you’ll get funneled into a potentially new corner of videos. With that in mind, I think Youtube recommendations are quite good. Just do it when your recommendations get stale, monthly or so. I wish other platforms had that like Amazon.

> 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".

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

#50

Post by zeroonetwothree »

ars wrote: post #26
sph wrote: post #17

Earlier quoted context omitted.

YouTube absolutely IGNORES your preferences. 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 click "don't recommend this channel", or "I don't like this video " every single day, I have an extension to make this conveni…

For me it recommends exactly what I watch - I've never seen any recommendation for stuff I don't watch. The trouble seems to be it can't seem to find anything new. So if anything, I'd want it to branch out a bit and find semi-related stuff, but it doesn't. You're not the first person to complain about YouTube recommendations, and I don't know why my experience is so different than yours.

My experience is the same as yours. I wonder if there’s some giant A/B test going on.
Post reply on HN