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

241–250 of 311 posts

Re: Continuing our work to improve recommendations on YouTube

#241

Post by paulddraper »

elihu wrote: post #225
paulddraper wrote: post #203

Earlier quoted context omitted.

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.)

I think the problem here is that "recommendations that maximize expected marketing revenue" and "recommendations that maximize user experience" are not the same. Most of us would prefer Google to present us the latter, but they have considerable financial incentives to present the former instead.

> "recommendations that maximize expected marketing revenue" and "recommendations that maximize user experience" are not the same

That, I think, hits the nail on the head.

I think it explains the Amazon case quite well. I'm not so sure sure about the YouTube idiocy.

Re: Continuing our work to improve recommendations on YouTube

#243

Post by AgentME »

aphextron wrote: post #226

I don't want YouTube to reccomend anything for me. I don't want Facebook to show me the news stories they think I should see. I don't want Google to show me the ads they think are more relevant to me. None of this improves my life in any way whatsoever, and in fact is leading to the dumbing down of society. We have completely handed over individual taste, interest, and curiosity to algorithms that do nothing but opti…

The users of YouTube, Facebook, and Google generally like a feed of things recommended to them. I feel like if this were an article about too many people pissing in a pool, your comment is saying no one should want to swim. That's one way to not get urine on you, but plenty of people clearly want to swim and it doesn't help much to blame them.

Re: Continuing our work to improve recommendations on YouTube

#244

Post by fzeroracer »

fvdessen wrote: post #239
fzeroracer wrote: post #224

Earlier quoted context omitted.

If you think they're conceptually quite similar then I don't know what else to say except that you would just be objectively wrong. The difference between a conspiracy theory and a social critique is the personal nature of it. People that believe that the 'Jews' or that 'The Deep State' is responsible for everything believe that there are specific actors, people doing things explicitly to stop someone. When people co…

> You can't point to a similar statistic, study or what not that shows 'The Deep State' is specifically targeting Donald Trump It usually gets more personal at the extreme. People will show you stats that shows a disproportionate majority of federal employees supporting the Democratic party and infer that there is 'systemic bias', some will point at specific cases and make leaders of hidden conspiracy out of them. I'…

You seemingly avoid the core conceit here, which is making me think you're not actually here to argue in good faith.

Yes, there are going to be people that will take anything to an extreme. That does not make 'Systemic Racism' somehow a conspiracy theory, especially on the level of things like the 'Deep State'. Arguing otherwise is to me the highest delusion.

Re: Continuing our work to improve recommendations on YouTube

#245

Post by Scoundreller »

kolinko wrote: post #218
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.

Some other online store, but related :) I once bought crab meat in a can from an online store (hard to get it any other way in Poland). A banner for buying more canned crab meat kept haunting me for the next two years, in various internet locations :)

I guess that’s proof that consumers overpay for it.

Re: Continuing our work to improve recommendations on YouTube

#246

Post by platzer »

munificent wrote: post #214
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…

This is literally mentioned in the blog post. > More recently, people told us they were getting too many similar recommendations, like seeing endless cookie videos after watching just one recipe for snickerdoodles.

Just checked my feed of recommended videos and it is still full of Tetris and excavator videos that I was into lately. It would be awesome to be surprised with some extraordinary lectures or documentaries similar to stuff I watched 5 years ago, or so.

Re: Continuing our work to improve recommendations on YouTube

#247

Post by paulddraper »

cycrutchfield wrote: post #189
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…

Read the Youtube recommendation paper ( https://ai.google/research/pubs/pub45530 ) and it will become very clear why it recommends a ton of high-engagement clickbaity content based off of a minimal set of recent watches.

Could you give a hint? I'm not seeing it.

Re: Continuing our work to improve recommendations on YouTube

#248

Post by AgentME »

dilap wrote: post #207

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).

>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".

If your local library or city center decided to prohibit a local white supremacy group from putting up flyers on their billboards recommending Holocaust conspiracy theory materials, would you criticize them the same way? I'm curious about why people seem to have very different standards for public spaces and websites moderating themselves.

Re: Continuing our work to improve recommendations on YouTube

#249

Post by scrollaway »

propogandist wrote: post #211
JaimeThompson wrote: post #130

Earlier quoted context omitted.

I think that videos that say kids didn't die in school shootings should not be allowed to make money off of said videos especially when they try to get people riled up to harass the survivors and their parents. Furthermore I expect Google to apply their policies evenly which they don't. If violations of Google ToS are reported they should be taken care of don't you think. It isn't like you can actually debate them on…

Youtube has made significant efforts to not show ads on content that is deemed unsafe, so those videos generally aren't making any money. You seem to be against people putting out content that don't align with your views, which is something completely different.

> You seem to be against people putting out content that don't align with your views

Not GP, but I see this a lot as a retort to people who make the claims they did and I tend to call BS on it.

I strongly suspect there is an extremely broad library of "content that doesn't align with their views", including an extremely broad spectrum of political views, they would have zero issues with on YouTube.

Somehow, climate denialism, antivax content (and other health hazards), holocaust denial, general nazi shit etc are consistently not part of that.

Now GP can come in, correct me and tell me that they want everything they ever would disagree with off Youtube for good. But somehow, I suspect they won't.

So, tangential to your post but, tell me: why is it that there's such a consistently-easy-to-define line for content that a lot of people think should be kept off various platforms, and why the hell is it that, whenever I see someone defending such content, it's almost always someone who belongs in the category of people who believe in said content, rather than a staunch defender of free speech.

I ask this knowing full well there are many people I hold in very high esteem, who legitimately do defend extremely vile shit they disagree with, based on free speech principles. Most of those people I personally know work at the EFF and I've never heard them say much about deplatforming nazis. (could it be because free speech is a government thing, not a youtube thing …)

Re: Continuing our work to improve recommendations on YouTube

#250

Post by sjwright »

rapind wrote: post #170
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…

I have the opposite problem. I'm only interested in new videos for the most part, and my recommendations are filled with videos from months ago. It would be much better if they would give us a flag or some date range options. Obviously this increases UI complexity, but maybe it doesn't need to be that prominent?

I think when the GP says new, he meant newest entry in their history. His problem is that watching one unusual video immediately and dramatically biases the recommendation engine at the expense of videos that match his long term viewing habits.

What you’re after already exists, the “recently uploaded” section on the home page.

Post reply on HN