Hacker News Reader

Comment by asciimov | Hacker News Reader

This is a damned shame. I use dislikes to gauge if a video is relevant to the topic at hand.

For example, I like fixing stuff and use Youtube for unfamiliar items. Often videos with a high number of dislikes are Spam (usually a bot reading text from some website), Videos where the person handwaves through the steps, videos where the fix is improper, or videos where the creator spends more time monetizing than working through the problem.

View on HN

Same thoughts I had, the more niche the topic the more important this feature was for finding a workable solution. A year back I was fixing a microphone and 9 of the 10 videos I looked at on it were convoluted expensive or ineffective work-arounds while one actually showed how to fix the issue relatively easily.

I'm really gonna miss this feature.

Another example of this: pronunciation videos.

Downvotes are a massive indicator if a chosen video is inaccurate.

Which is kind of funny when you think about it. Who would watch a pronunciation video if they already know how it’s pronounced?

Not discounting what you said. Just found it amusing :)

I suppose it could be people checking multiple sources and coming back to downvote after assessing which is the right one. At least that's the world I want to live in. A guy can dream.

> people checking multiple sources and coming back to downvote after assessing which is the right one

I do this on stackoverflow. I wouldn't be surprised if this is what happened here as well.

Often time when I look up a pronunciation video, it's for a language I'm learning and I have a vague idea of how the word is supposed to sound, but not sure on things like tonality and stress.

So, if I were to look up how to say "ramen", I might not know if it's supposed to be ra-MEN or RA-men, but I'd know it's wrong if they say it like "r-amen".

Some heavily downvoted videos are great.

Here's a classic: https://www.youtube.com/watch?v=SXmv8quf_xM

Currently rated 3.7k likes, 15k dislikes. It's so good.

In the past I have used dislikes to tell if chain sawing tutorials are good or if the operator is insane. It works well for things like that.

That's exactly when dislikes are so useful. This screams of a bad/easy product decision by someone who doesn't use YouTube this way. Facebook, ebay, etc went to this so I guess that's the future. No negative reviews to maximize time-on-site.

For recommended videos I see only two paradigms

- videos with >95% likes

- mostly political videos that have obviously been brigaded

imo this is a reasonable change

Really?

At least when you search for a video topic, you are likely to get a wide range in quality. Regularly surfaced content has a variable amount of dislikes, and almost always the dislike count is signal and not noise.

I rarely engage with YT recommended videos so I can't verify the accuracy of your claims. Either way, however, it sounds like reworking their recommendations algorithm would be a more reasonable way to deal with this issue than simply hiding the like/dislike ratio. How is this change going to meaningfully affect which videos show up as recommended?

Seems like a combination of top-rated comment and likes divided by viewcount will yield similar information, though...

Big assumption, have you run the data? I doubt it would be a good proxy.

Make a video on a taboo topic. During the first wave you end up with tribal zealot commenting and liking with low/medium viewership.

This pushes the video in the mainstream which causes more views but smaller likes counts. The top comments increase with flamewars in the comments.

Without the dislike the popularity could be negative.

With this move I expect non-positive comments days to be numbered. Any feature that doesn't align with brand positivity is fair game to be removed.

Reply on news.ycombinator.com