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

Continuing our work to improve recommendations on YouTube

youtube.googleblog.com

Post reply on HN
Print view View original

201–210 of 311 posts

Re: Continuing our work to improve recommendations on YouTube

#201

Post by rasz »

Ivoirians wrote: post #110
atonse wrote: post #40

We currently don't allow our son to watch anything on youtube, simply because it's a cesspool of horrible videos for children. You're always almost one click away from some violent video where cartoon characters are getting their limbs chopped off, etc. And it's sickening that Google's done nothing about it. Secondly, I can't watch any video without it taking over my recommendations. Alex Jones for example. I still h…

If you delete a video from your Youtube watch history, it doesn't get used for recommendations anymore. https://support.google.com/youtube/answer/95725?hl=en

Its a Lie. I have YT watch history disabled for hmm 2 years now, recommendations "work" just the same.

Re: Continuing our work to improve recommendations on YouTube

#202

Post by FiveSquared »

Elitroness wrote: post #18
FiveSquared wrote: post #5

Earlier quoted context omitted.

However, who determines what is a conspaircy? The government might be deciding that Iraq has WMDs and any evidence to the contrary is a conspiracy. While some measures are needed to be taken, who decides what is? No one is truly impartial. The censorship (while required due to how much crap is in YouTube) might devolve to 1984 without comple transparency, which we all know google isn’t providing.

Who determines when daytime is? is 7AM daytime? is 7AM in Alaska in the middle of winter daytime? There is a discussion to be had about some topics at the edge...but 2PM everyone agrees is part of daytime; the flat earth 'conspiracy' should never be recommended to anybody. And sadly, in the current climate, this also has to be said: This does not mean Google is picking sides. If you on your own reflections decide tha…

Truth, but however, the road to hell is paved with good intentions.

Re: Continuing our work to improve recommendations on YouTube

#203

Post by paulddraper »

maxxxxx wrote: post #162
mindcrime wrote: post #145

My biggest problem with YT recommendations is that they dramatically overweight recent videos when generating the recommendations. I mean, I can spend a week doing nothing but watching (well, listening) to videos of classical music performances, and my recommendations will be full of classical music, exactly as you'd expect. Then I watch one random one-off thing, like "15 funny pitbull fails" and suddenly ALL of my r…

All the AI stuff is really still incredibly dumb. Same for Amazon: Buy a mixer and suddenly you will be haunted everywhere by ads offering more mixers.

You think that's dumb, but I bet the marketing data behind that says otherwise.

You think "I already have a mixer, idiot!" but in reality the chance that you are interested in buying a mixer just went from 1% to 2%. (Because you want another one, because you want to gift one, because you returned yours, etc.)

Re: Continuing our work to improve recommendations on YouTube

#204

Post by rasz »

alttab wrote: post #127
atonse wrote: post #116

Earlier quoted context omitted.

Even if you logout, your anonymous session has its own set of recommendations. No?

Disable cookies

you will get served "interests shared by users from that particular IP", google helpfully assuming its a single household.

Re: Continuing our work to improve recommendations on YouTube

#205

Post by AgentME »

throwawaysea wrote: post #193
AgentME wrote: post #161

Earlier quoted context omitted.

Maybe they should be there for people who seek them out, but YouTube shouldn't be recommending them onto people who didn't previously want them and who might not be approaching it skeptically.

Why is it YouTube's business to judge truth and falsehoods and provide a manicured walled garden? That seems like an avenue for all sorts of manipulation, censorship, etc.

YouTube is already an avenue for all sorts of manipulation. They recommend harmful crazy conspiracy videos to unsuspecting people all the time right now. (9/11 conspiracy theory videos might just sound silly, but I'm also thinking of stuff like racially-charged holocaust conspiracy theories and baseless anti-vaccine stuff.) I think you're way too optimistic if you don't think those are likely converting many people at YouTube's scale.

Just because this is the result of non-action by YouTube doesn't make it okay. YouTube becoming a moral arbiter isn't a perfect situation, but it can easily be better than the status quo.

Re: Continuing our work to improve recommendations on YouTube

#206

Post by aasasd »

jalgos_eminator wrote: post #184
aasasd wrote: post #140

Earlier quoted context omitted.

> an artist from Belarus that makes fantastic synthpop Alright, I'm gonna need a link.

https://www.youtube.com/watch?v=ypknY6nIDe0 Check the comments. It looks like there were plenty of people who had the same experience as me.

Ah, I'm not sure that this recommendation is tied to user's history of watched videos. Because I've also seen it in my ‘recommendations’―and though I tend to occasionally listen to a couple tracks of Gary Numan or a vaporwave mix, I don't think a 2018 release would follow from that dataset.

Re: Continuing our work to improve recommendations on YouTube

#207

Post by dilap »

honestly, i'd like to be the judge of if a theory is crackpot or not. i'm not a kid, i don't need youtube to save me from "bad information".

hopefully the change is mild. i've discovered a ton of great stuff through youtube recommendations (especially music! youtube blows everything else away for music discovery).

Re: Continuing our work to improve recommendations on YouTube

#208

Post by fvdessen »

fzeroracer wrote: post #75
jlawson wrote: post #51

Earlier quoted context omitted.

This doesn't match reality. I've watched lots of conservative videos and I've never seen anything about flat Earth or "insane conspiracy theories". Honestly this comment thread seems like trying to build a narrative bridge between straightforward right-leaning content and bizarre conspiracies. The goal is to use censorship of bizarre conspiracies to justify censoring right-leaning content by conflating the two. I'm v…

Conspiracy theories are, for the most part, a right-wing thing. Unless you can give me examples of conspiracy theories and their peddlers on level of Alex Jones for the far left.

GMO/Monsanto, Big Petrol & all their wars, Putin controls XYZ, "The Oligarchy", "The Patriarchy", "The Privileged", Plastic straws from Europe killing turtles in the Galapagos, "Systemic Racism", The Gender Gap, etc.

Re: Continuing our work to improve recommendations on YouTube

#209

Post by slig »

EpicEng wrote: post #182
mindcrime wrote: post #145

My biggest problem with YT recommendations is that they dramatically overweight recent videos when generating the recommendations. I mean, I can spend a week doing nothing but watching (well, listening) to videos of classical music performances, and my recommendations will be full of classical music, exactly as you'd expect. Then I watch one random one-off thing, like "15 funny pitbull fails" and suddenly ALL of my r…

It's terrible. I have to completely log out when I watch videos which put me to sleep a couple times /week because I don't want my suggestions filled with dry history lectures.

If you're on mobile, there's an incognito mode (at least on Android).

Re: Continuing our work to improve recommendations on YouTube

#210

Post by maxxxxx »

paulddraper wrote: post #203
maxxxxx wrote: post #162

Earlier quoted context omitted.

All the AI stuff is really still incredibly dumb. Same for Amazon: Buy a mixer and suddenly you will be haunted everywhere by ads offering more mixers.

You think that's dumb, but I bet the marketing data behind that says otherwise. You think "I already have a mixer, idiot!" but in reality the chance that you are interested in buying a mixer just went from 1% to 2%. (Because you want another one, because you want to gift one, because you returned yours, etc.)

The mixer probability maybe went from 1% to 2% but there must be 100s of other products they could show that have much higher probability. Maybe a mixing bowl or a cookbook?
Post reply on HN