Live data from Hacker News

Mozilla Is Crowdsourcing Research into YouTube Recommendations

foundation.mozilla.org

121–130 of 179 posts

Re: Mozilla Is Crowdsourcing Research into YouTube Recommendations

#121
post #62

Couldn't care less about this. Just put out a paid version of Firefox that dedicates this revenue stream exclusively to continued Firefox development and I will happily pay fist fulls of money for it. Seriously, Firefox adds so much value to my life and I'd happily pay for it.

https://donate.mozilla.org/en-US/

I think the "exclusively for Firefox development" is pretty important. Mozilla is doing a lot of random things these days

Re: Mozilla Is Crowdsourcing Research into YouTube Recommendations

#122
post #38

Earlier quoted context omitted.

I use YouTube infrequently. Usually for binging random content like comedy sketches or speedrunning. No matter what I'm on, there's always a few videos in my recommendations that make no sense and look like trash. Not generically, but specific videos that ALWAYS appear. One of them is "I PAID FIVE BASSISTS ON FIVR TO PLAY AN IMPOSSIBLE BASSLINE" ... Why is this so bad? What are your hundreds of data scientists doing?

They stop appearing if you specifically mark them as not interested

I often AM interested in a topic, I just don't want every little piece of content drivel shoved down my throat. I still want the cream of the crop to occasionally pop up in my feed.

That's the problem with marking things as not-interesting. The algorithm will wildly overreact in the opposite direction.

Re: Mozilla Is Crowdsourcing Research into YouTube Recommendations

#123
post #105
post #91

Earlier quoted context omitted.

Literally the first sentence of the article is about dangerous recommendations. And the second is about harmful content. This is not just an open-ended inquiry of what youtube is recommending. It's about finding the bad and getting rid of it. EDIT: I also want to remind people about how Mozilla used its push notifications to call for a Facebook boycott over objectionable content.

> YouTube recommendations can be delightful, but they can also be dangerous. The platform has a history of recommending harmful content — from pandemic conspiracies to political disinformation — to its users, even if they’ve previously viewed harmless content. What part of this is untrue? And again, you're saying "getting rid of it" without evidence. It's a study. It does not mention banning anything, anywhere. If th…

> What part of this is untrue?

The characterization of pandemic conspiracies and political disinformation as harmful.

It's like kids and allergies. You never hear about kids who grew up on farms being allergic to animals - it's always those whose parents didn't have any when they were growing up.

If people aren't subject to misinformation, they'll never develop the sense of who's lying and who isn't.

It used to be that we gave common-sense advice - "don't believe everything you read on the Internet". Now, it's the other way around - "we must cleanse the Internet of harmful content".

Being exposed to misinformation is good for you, and it's good for democracy.

Re: Mozilla Is Crowdsourcing Research into YouTube Recommendations

#124
Am I incorrect in assuming that Mozilla’s goal was at one time to expand access of information to the world?

I don’t see any other likely outcome of this “research” other than a fresh new media hue and cry detailing chains of “wrong-think” that will lead to activists who masquerade as journalists calling for more content they deem unfit for the plebs consumption.

Re: Mozilla Is Crowdsourcing Research into YouTube Recommendations

#125
post #53

Earlier quoted context omitted.

I use YouTube infrequently. Usually for binging random content like comedy sketches or speedrunning. No matter what I'm on, there's always a few videos in my recommendations that make no sense and look like trash. Not generically, but specific videos that ALWAYS appear. One of them is "I PAID FIVE BASSISTS ON FIVR TO PLAY AN IMPOSSIBLE BASSLINE" ... Why is this so bad? What are your hundreds of data scientists doing?

I'd say there's definitely a lot of "LOOK AT MY YOUTUBE SEO TITLE" but in actuality that bassist guy is really good and generally plays some great riffs.

It's strange.

It's like a wild species with vibrant, sometimes garrish colors saying 'Look at me!', where to discriminant viewers it's typically poisonous.

I'm pretty cautious.

But for a large number of my favorite content creators it's a bait and switch. Bait your clicks and actually provide really meaty content.

That aside, my personal favorite trend in the digital media space are the 1-4 minute tutorials for people who already know the software.

These people typically make money outside of Youtube and don't rely on snaring views.

Re: Mozilla Is Crowdsourcing Research into YouTube Recommendations

#126

Earlier quoted context omitted.

I use YouTube infrequently. Usually for binging random content like comedy sketches or speedrunning. No matter what I'm on, there's always a few videos in my recommendations that make no sense and look like trash. Not generically, but specific videos that ALWAYS appear. One of them is "I PAID FIVE BASSISTS ON FIVR TO PLAY AN IMPOSSIBLE BASSLINE" ... Why is this so bad? What are your hundreds of data scientists doing?

The function of an adtech recommendation system is not to find content that you want to see but rather to show you content that's the most profitable for the adtech firm that's serving it. The videos that YT recommends over and over again are probably videos that have a high ratio of advertising revenue (meaning good demographics) to playback cost (meaning low overall byte size) and a high likelihood of going viral b…

Recommending you a video you don't want to watch is a wasted opportunity. If you don't want to watch it, you won't, so YouTube's earnings are zero. If something you do want to watch is recommended instead, you will, and YouTube's earnings are nonzero.

I'm right now on a browser without my account and on which I don't usually watch YT, basically only when someone pastes a link while gaming. About half of the front page are things I might have clicked on when bored. A third are long. Two are definitely controversial. I don't buy your hypothesis.

Disclosure: I work in Google, but my closest contact with YouTube was drinking whisky with a YouTube SRE some four years ago.

Re: Mozilla Is Crowdsourcing Research into YouTube Recommendations

#127
post #26

Earlier quoted context omitted.

For me it has been the opposite as in it has so little hysteresis. Just one day I watch something I do not usually watch, say cat videos, and for weeks I will get only cat videos as recommendations. I think their algorithm needs to be cognizant of and adapt to temporal shifts in a viewers interest. Some users may like change in focus at slightest of deviations, others may prefer more hysteresis.

Because of this, when I'm sent video links to watch I will open them up in a Private Window / Incognito so I do not pollute my account.

I set my browser to block all cookies. Has the same effect as private window mode. That way I can still browse youtube without having to open a new window. Yea, it sucks that I can no longer sign in, but at least the recommendations are better.

Re: Mozilla Is Crowdsourcing Research into YouTube Recommendations

#128

What's impressive to me is that YouTube Recommendation are so ... stale. All it recommends to me is content from the same 3 categories, and the same videos over and over and over (you would think if I didn't click on the suggestion in the past 6 months it would adjust). Now the topics it recommends to me are non-political, but I can very well see that if you fall down one of more extremist rabbit holes thats all you…

I honestly don't understand this either...

Youtube knows exactly which videos I've watched and which I haven't.

Why does it think it's a good idea to ever recommend videos that I've already watched?

Re: Mozilla Is Crowdsourcing Research into YouTube Recommendations

#129

What's impressive to me is that YouTube Recommendation are so ... stale. All it recommends to me is content from the same 3 categories, and the same videos over and over and over (you would think if I didn't click on the suggestion in the past 6 months it would adjust). Now the topics it recommends to me are non-political, but I can very well see that if you fall down one of more extremist rabbit holes thats all you…

[semi-speculation presented as if I actually know] Youtube recommendations are best effort. That effort has a computational budget and a time limit. The fastest and least computationally intensive recommendation is to pull the last set of recommendations out of the on-device local cache. The next fastest thing is to recommend whatever is cached near the edge of your internet connection whether it fits your previous v…

>Of course, this ignores the issue of Google revenue. Advertisers targeted you, target you based on the videos you've watched. So there's that too. Google recommends videos for which it has relevant advertisers.

I think this is exactly the problem.

Contrast this with the recommendations I get from my paid spotify account where I don't get ads: It's absolutely Excellent. Seriously, whoever worked on Spotify's recommender system algo deserves a huge pat on the back. They have the best recommender I've encountered anywhere on the net, and I pay a lot of attention to recommenders. It's just light years ahead of everyone else.

Re: Mozilla Is Crowdsourcing Research into YouTube Recommendations

#130
post #128

What's impressive to me is that YouTube Recommendation are so ... stale. All it recommends to me is content from the same 3 categories, and the same videos over and over and over (you would think if I didn't click on the suggestion in the past 6 months it would adjust). Now the topics it recommends to me are non-political, but I can very well see that if you fall down one of more extremist rabbit holes thats all you…

I honestly don't understand this either... Youtube knows exactly which videos I've watched and which I haven't. Why does it think it's a good idea to ever recommend videos that I've already watched?

I think there's a category or two where it makes sense:

- fitness videos

- music

- how-to & instructional videos

Post reply on HN