At some point Youtube search was neutered. I get 5-10 results for any given query before Youtube gives up and just offers me recommendations rather than search results. Interestingly, if I search google for [search term] site:youtube.com, I actually get pages of results which are never present in the Youtube search. Sadly, it seems that Google has intentionally diminished Youtube's own search function in order to prioritize use of the recommendation system.
It doesn't even have to be that recent in some cases. Searching for "Sandy Hook Elementary" along with just about any other search term is going to give you a page with results about a shooting that took place a decade ago. Searching for a proper noun when they are the top story of the day is pretty much only going to give you that one story.
Then they removed that value point and aligned what they presented with my existing bias - at which point I stopped using it.
As a user of them, I make sure to use multiple engines (preferably with distinct indices and jurisdictions) when I research something essential that affects my long-term future. This is because I distrust how politics is done (industry lobbying).
You may be interested in our alternative YouTube recommendations. Search a channel to find a list of other channels making similar videos. For example, our recommendations for How Money Works include other established econ focused channels like Economics Explained and Money & Macro but also smaller more obscure channels like Wall Street Millennial:
Likewise if the original image has already disappeared from the net – finding a copy via Pinterest is still better than finding nothing. (Interestingly Bing's search can serve a similar purpose – compared with Google, it feels like Bing is much more reluctant to remove 404'd images from its index again. While those broken results are annoying, too – and unlike Pinterest neither Google nor Bing store (at least publicly) full-resolution copies of those images – at least you can try plugging the preview image into reverse image search and with some luck maybe find an alternative copy of that image that has survived elsewhere.)
And finally, the keywords added by Pinterest are possibly also helpful in making images more discoverable. (This might be no longer that relevant now that Google is also running image recognition on the images it indexes and no longer has to solely rely on the hope that any text near that image or the rare "alt"/"title"-text is actually relevant to the image contents, but it might still help a little.)
Google is resting on its laurels - the perfect recipe for losing hegemony.
(you might have to change your region to usa first)
How management thinks it is a good idea to make their products useless by letting a bunch of self-aggrandizing "analysts" or marketers or whatever attempt to manipulate their customers/viewers is beyond me. The fact that they are all paid to fork it up so badly is just mindboggling.
Maybe I'm out of touch but I don't think this is an unreasonable expectation.
One trick to get rid of them is to select one/any additional search filter (for example select HD only).
It is absolutely ridiculous that you have to do this, but it works for now.
I search from the command line using a short shell script.^1 Perhaps it depends on the query, but I consistently get more than 5-10 results. For example, I might get 26 reults with the initial query. By continuing the query repeatedly I can usually get over 1000 results total. YouTube search results are delivered as JSON so I wrote some simple, quick and dirty utilities in C to reformat the JSON to TSV with only the details I want. The workflow and example script below use those utilities, i.e., yy059, yy062 and yy064.
# YouTube search workflow:
# initial query
1.sh query > results-file
yy064 < results-file|yy062 > results.tsv
# continuation
1.sh < results-file >> results-file
yy064 < results-file|yy062 >> results.tsv
#!/bin/sh
HOST=www.youtube.com;
case $# in :)
;;0)
KEY=AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8;
MIN=\{\"context\":{\"client\":{\"hl\":\"en-US\",\"gl\":\"US\",\"clientName\":\"MWEB\",\"clientVersion\":\"2.20210916.06.01\"}},\"continuation\":;
TOKEN=$(yy059|sed -n "s/$/}/;s/\"token\":/$MIN/p"|tail -1);
URL=https://$HOST/youtubei/v1/search?key=$KEY
curl --http1.1 -A "" -H "Content-Type: application/json" -d "$TOKEN" "$URL"
;;*)
S=$(echo "$@"|sed 's/ /+/g');
URL=https://$HOST/results?search_query=$S;
curl --http1.1 -A "" "$URL"
esac;
Absolutely agree that YouTube search got neutered.
youtube.com##ytd-search ytd-item-section-renderer>#contents>:not(ytd-video-renderer),ytd-secondary-search-container-renderer
youtube.com##ytd-search ytd-video-renderer:has(.badge:has-text(/new/i))
youtube.com##ytd-watch-next-secondary-results-renderer ytd-compact-video-renderer:has(.badge:has-text(/new/i))I suspect it might be supplying specific filters (e.g., "intitle", mentioned elsewhere in this thread) to YouTube's query engine.
https://github.com/TeamPiped/Piped/blob/860ca92a4d49a85a3781...
That looks strongly like a simple pass-through.
An unqualified search reports (at this writing) just north of 30 million results (30,000,229). It will display only 34 pages' worth of results, I believe 1,000.
See: <https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...> Final page: <https://hn.algolia.com/?dateRange=all&page=33&prefix=false&q...>
As discussed recently on HN, the match count is useful information even where all matching results are not displayed as it indicates whether or not a query is generic.
Contrast search: <https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...> (one result as of this writing).
Edit: Add to that the increase in number of ads per video and their length...ugh
Kinda think there are more than 10 videos about a major food dish on YT.
Starting to think the backend has a concept of “blessed” channels to surface
That's because recommended videos are more popular, and more popular videos earn them more ad revenue.
Now consistently I get adverts from Facebook for that same product from other scammers, no matter how much I block them. Presumably because that advert 'engaged' me for the longest. Eugh.
Incidentally, ASA responded that it was not a UK company ... which sounds like "we don't care about scam advertising we're just here for the paycheck". The company were advertising on UK Facebook and held a UK registered trade mark.
• I've conceded that adverts are going to show up in my feed regardless and most advertisers can't be bothered to target properly so are mostly irrelevant
• I take great pleasure in "engaging" with the obvious scams and shitty crypto/NFT ads in the hope it's eating their ad budget
• this makes the adverts I see extremely irrelevant in a reinforced feedback loop
- search is neutered/broken, so finding new interesting videos is a chore
- recommendations need to be manually combed regularly to remove dumb videos
- watch history also needs to be combed to remove potential false signals
That's a lot of curation to get somewhat decent recommendations. Not saying we have a choice, it's not like Vimeo will make a comeback to save us, but that's a pretty bad place we're in.
It, unsurprisingly, has a history of _lots_ of different things, you'll be shocked about how much Google knows about you. Probably things you yourself have forgotten.
It seems like the problem of asymmetry in recommendation platforms is more a choice from a design perspective than a challenge yet to be figured out.
I'm far from an expert tho so just a thought from a laymen
For them to sell user experience they'd need to have competition breathing down their necks first
But there was a time where it seemed like it could be what the people deserved, even if it was just teenage optimism lol
I wish I could somehow "save" the state of my recommendations profile so I can revert it if it goes off the rails.
Low quality videos also get nuked with prejudice from my history.
this and it's going too far, since Youtube is now polluting people's search with very questionable recommendations. For instance, I'm into CGI and Youtube search kept recommending me some random footage about a female streamer which behaved and dressed very suggestively, that already had millions of view, what does it have to do with Shaders and Raytracing? Why the hell does Youtube want me to look at that?
Think how it must feel to be profiled to be a crook or a bad credit risk or something else that actually impacts your life and is false. It's a little worrying.
[1] Other people like you like this so it is assumed you do also. Without "lecherous" ever being mentioned or even anything like suggestively dressed appearing as a model feature. It is really, really easy to accidentally train a statistical model to be wildly prejudiced. It seems significantly harder for us to categorize it when it happens as "other similar decisions are made by shockingly violent racist and evil humans and this model and those using it should be treated with the same or more derision as they claim not be so and ought to know better." Significantly harder for many good reasons as well as bad ones.
But that isn't happening, and won't ever, right? Right?
"A modern clinician needs to remain abreast of social media sources as knowledge of media content is essential in managing patients in the current environment."
https://movementdisorders.onlinelibrary.wiley.com/doi/abs/10...
We've all met people that adopted wholesale the mannerists of this-or-that online influencer; gestures, tonality, jokes and all. It's creepy as hell to see, and sure, and algorithms are reinforcing this phenomenon. An "obsessed fan" is YouTube's ideal viewer, since his engagement will be through the roof.
Spotify is unprofitable, so they allow the biggest artists to pay to be promoted by their algorithms, which shapes the world's music tastes.
Fat people are being recommended mukbang videos.
These algorithms can be studied, as TFA does, and you can try to hold these companies accountable, but there's limited utility in doing scientific studies of algorithms that get updated multiple times a day.
In today's world, manmade algorithms from a handful of huge tech companies are likely exerting more influence over human evolution than nature itself is!
One answer is that some assumptions that YouTube makes about the viewers is based on very broad stereotypes; the geek who’s into CGI also has a sexual side that they are about to discover through YouTube.
I don't like YT quite honestly, and it's been a while. I understand it's a business but you don't have to provide such bad experience to make money.
For searching videos I'm more and more relying external search because with YT I just get suggested crap, like with google. I'm beyond tired with this.
I think offering alternative titles for each language is on the creators, not youtube themselves
This was the last drop for me, when I started to use external search.
Similar story with a guy who reviews indie games daily. I want to see the future (potential) indie hits not the games from years ago.
The most blatant was a specific "Get robux" video that appeared constantly on my feed from dozens of different channels but that triggered this extra engagement in the form of blocking, reporting and disliing
Click triple dot thingy on recommendations and select "not interested"
Once YouTube think's it knows what I like, I get 1 or 2 categories of videos I might watch and barley anything else related to my subscriptions.
If I had to describe YouTube's recommendation algorithm I'd say "undefined behaviour". It might recommend the same uninteresting video every day, but also hide an interesting, subscribed channel's videos for months even though I liked almost every one of their videos. It might mix videos I already watched once with completely unrelated search results. At this point Google could probably improve their algorithm just by flipping a coin.
YouTube is hilariously bad.
It doesn't matter how many times I say I'm not interested in K-pop, it doesn't help.
So far I’ve found it does stop showing you if you remove what caused it from history.
Unfortunately, Capitalism doesn't work like that :-(. Back in the old days, Google searches used to be organic searches in the genuine sense, and that was the best Google we ever had. But today's Google is the Big Tech Capitalist and so will be their algorithms. The organic logic should follow your prunes and show you what you want but the capitalist logic will only show what they're paid to show to you! (i.e. propaganda, politically motivated content, inciting content, etc. etc.).
On your other point: I've not read anything that suggests Google is paid things to promote content. Where do you get that from?
If I click on anything I wouldn't usually watch, it is then the only thing that goes in my recommendation for a good 2-3 weeks. I watched one dashcam video the other day that someone linked me, which had the tags Tesla and Elon Musk.... Now the only thing I get in my recommended is incredibly strange videos about Elon Musk and Teslas.
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?
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've been watching lately.
Personally, I'd love it if you could curate the recommendation system a bit more. I'm never interested in rewatching stuff I've already seen. And the recency bias for recommendations means you're always getting more if what you've recently consumed. I'd like it if more older interests were mixed in. Finally, I'd love to be able to blacklist certain topics and games entirely.
There's so much missed potential with YouTube recommendations.
Some might say that you can still use search, but that has been crippled to include unrelated recommendations after one page of results. Not only that, I want YouTube to help me discover new content. When the recommendation system works and it shows me new kinds of interesting stuff, it's superb. TikTok is eating YouTube's lunch in this regard.
Edit: forgot to mention this. On mobile if you do a recommendation refresh it'll show a New to You filter which you can use to explore more diverse recommendations. I don't know why it's hidden in this manner.