please empty your brain below

There's a theory that the robot uprising won't be through a war of bloodshed. Instead, we'll eventually create algorithms to recommend everything in our lives to us, and we'll end up being too lazy or inept to choose anything different to what the algorithm says, as it's easiest to just go with the flow. Ergo, slowly but surely the algorithms end up having total decisive control over our lives.

In YouTube's case, many creators have been screwed over by The Algorithm's insistence in preferring specific types of content to recommend - usually inane, bitesize (10-15 minutes) - such that their own content doesn't get nearly as much exposure. YT's bosses can do very little about it - the Algorithm uses machine learning, so it learns by itself, so basically no-one actually knows how it works at this point.
Try DoggCatcher for podcasts, it gives lots of control
Thanks for your recommended blogs that I may have missed or not previously caught up on. Most helpful
I'd say its impossible to be shown everything - how often do you get any YouTube recommendations from a non-English speaking creator?

I do find YouTube works for some music stuff - for example something non-genre specific like 12'' singles throws up all sorts of gems, that led on to 'off grid' remixes, but you wouldn't find them looking for established artists - you'll just get recommended more established artists.
Amazon used to have a feature that allowed you to exclude viewed or purchased items from affecting your recommendations but I haven’t been able to find it for a couple of years. I wish that they would reinstate it and others would adopt it.
We press the 'subscribe' button and hey presto we get no notifications from those channels but still the algorithms' choices of useless, uninteresting, blah fodder.
As much as algorithmic recommendations can be deeply annoying - particularly the auto-playing kind - I suppose that, when I think of them in context of what came before (limited/no choice), they're probably a much lesser evil. Yesterday I listened to my hand-picked playlist, a random selection of tracks from my music library, a livestream of a radio news programme from a different country, a podcast recorded a couple of hours earlier, and watched an episode of a new US TV show while working, on several different devices, at the office, on the road, and at my home. Young me would not have believed that amount of choice and flexibility existed, as dg's penultimate paragraph suggests.
Nearly all services with recommendations let you edit your history, Youtube and Netflix certainly do. Youtube lets you turn off auto play suggestions, though you have to set this again on a Roku.

@Medford Huxley's Brave New World?
Youtube recommendations are turned on and auto play is turned off and I find this the best way. Usually the recommendations are fairly relevant and have the occasional gem that leads me off to new discoveries.

I think the most irrelevant is where a site actually states "A person who bought X also bought Y and Z". Total bxx. I wonder how many buyers fall for this and add the items to their basket whether they actually need them or not.
Oh come now. The "A person who bought X also bought Y and Z" is often unintentionally hilarious. I derive my amusement from the smallest things, sometimes. Eg: "People who bought balaclava also bought aluminium baseball bat and bear repellent"
My favourite is when you buy something from a website, and it continues to recommend it (or a similar product) despite the fact that the item is clearly not scalable for the average person. "Another dishwasher to put next to the one I bought from you last month? Don't mind if I do!"
You can turn off Recommendations on the BBC by going to your account settings and turning off personalisation. It does blatt ALL personalisation though.
I use YouTube like the radio at work, leaving it playing in the background. Its recommendations are far from perfect, but they're better than any I've got from a human - and unlike a human DJ, I can skip tracks I don't like. As the joke goes, a DJ is already an incomprehensible neural network who picks things based on who knows what.

When I care enough, I'll pick the next song myself. But most of the time, I don't.
once you see how inept BBC iPlayer is at recommendations, you realise how good YouTube is !

i find going into my Watched history and deleting items is the best way of stopping YouTube pushing similar videos.
Goes right back to the development of word processors and other office software. The number of times I've ended up shouting at computers "Stop trying to be 'helpful' and do what you're damn well told!"
I share some of DG's dislike of algorithms recommending content. I do my best to ignore the BBC's ones and I actively dislike them on Spotify.

In slight contradiction to this, in my day job I am responsible for the recommendations algorithms on a fairly well known retail platform. I see the engagement data and I have an understanding of how recommendations drive (and don't drive) behaviours.

I've come to the conclusion that the most important thing a platform can do to make its recommendations useful is to be transparent about how and why it's recommending something. Amazon does this pretty well - for example "people who viewed this page went on to buy" is directly useful, informative and encourages people to spend. The BBC (and others) do this badly.
Try leaving YouTube running all night and see what pit of hell it's descended into when you wake up in the morning. The good thing is it'll now think you chose to watch those videos and you'll never get rid of them.
If you subscribe to particular channels and then set up a shortcut in your web browser to youtube.com/feed/subscriptions you ONLY get stuff you've subscribed to.
Ad blockers seem to help too.
Make your own 'mixtape'.










TridentScan | Privacy Policy