4 ms·

Algorithmic feeds are always hostile to the user. They limit what a user sees, and slowly infantilizes them by teaching them to passively consume rather activel

Algorithmic feeds are always hostile to the user. They limit what a user sees, and slowly infantilizes them by teaching them to passively consume rather actively seek information. I've never encountered a situation where an algorithmic feed made things better. At best, they make things more addictive.

charcircuit 5y ago
>They limit what a user sees No, a screen has limited real estate. There is only so many things that can be shown to you at once. So on youtube for example there maybe only be room to show 8 videos. Ideally I should find all 8 videos interesting. If I don't, that's a waste of space and a waste of my time. Algorithmic feeds serve to improve the results such that I am more likely to watch a video that is put in one of those 8 spots compared to if I was just shown trending videos or videos from my subscription box. It is not hostile to me because it saves me time in finding something to watch. You are not being turned into a baby just because you no longer have to search for stuff manually. There is a difference between wanting to find something specific and just wanting to find something interesting. >I've never encountered a situation where an algorithmic feed made things better. One of the best algorithmic feeds is TikTok. After that YouTube's is pretty good to. I suggest trying one of those sites out to just see how effective they are in finding content that interests you. I have found algorithmic feeds have at least made things net better. There is too much information in the world to have to search through all of it ourselves. These algorithms solve the problem of dealing with this large scale of data.
- epicide 5y ago
  > Algorithmic feeds serve to improve the results such that I am more likely to watch a video that is put in one of those 8 spots compared to if I was just shown trending videos or videos from my subscription box. Right, that's the promise from algorithmic feeds. Sometimes, for some time, it's even the reality! I think there's some middle ground and nuance to be had here, though. I agree that just because there's "an algorithm" doesn't mean it's bad or a worse user experience. Finding interesting videos is a great example. However, I think it's fair to also say that video recommendation algorithms have gotten out of hand. We have a lot of content creators who are influenced in how to make their videos by what is popular. After all, if I make a video that is obviously more successful than another one, that means my audience likes it more, right? What did that video do differently that made it so much more successful? After a long enough time, you end up inadvertently (or intentionally) catering to the whims of the algorithm and not just your audience. So the issue I have with algorithmic feeds isn't that they are outright a bad idea for users. The problem is that it is not just a feature for users. It starts to shape the content on the platform. This can be used to subtly influence content creators to make better content, but without any idea of what goes into these algorithms, we are left to make wild religious guesses. And we have to assume that these algorithms serve the (financial) needs of the companies behind them and their sponsors, first and foremost. If not now, then gradually, by the same feedback loop as happens to the content creators.
- lukevp 5y ago
  The problem with this is that your suggestions are based on past viewing history, which isn’t always related to future needs. Examples: on one youtube account we primarily watch different types of media on, we watched a Labrinth music video. Neither my wife or I like labyrinth, but now our music section in YouTube is completely consumed by labyrinth music videos. One or two viewings isn’t indicative of us being labrinth fans. This is what we mean by the AI is dumb and makes things worse for the user. The music video section on that account is basically useless now. The same thing will happen with anything - one time I watched a restoration video for an old lamp I had, and YouTube assumed I now mostly wanted to watch antique restoration videos for the next 6 months.
  - zimpenfish 5y ago
    Same issue here - if I watch one video from a channel I'm not subscribed to, my recommendations will be flooded with that and related channels. Even picking "don't recommend this channel" only cuts down the spam, doesn't stop it - which is annoying because, sometimes, I've enjoyed what I've watched and I wouldn't mind the odd recommendation ... just not a dashboard full every single day from then on. Oh and "XYZ viewers also watch... [5 videos by XYZ]" is such an egregious failure that I can't help but laugh every time I see it.
  - eloisius 5y ago
    God. You watch one ICP video so that you can laugh at the “fucking magnets, how do they work?” line…
version_five 5y ago
Worse, they incentivize "creators" to post "refrigerator door wisdom" type stuff optimized for likes or clicks or whatever (or outrage-bait, but that seems to be more in non-professional forums).

charcircuit 5y ago

>They limit what a user sees No, a screen has limited real estate. There is only so many things that can be shown to you at once. So on youtube for example there maybe only be room to show 8 videos. Ideally I should find all 8 videos interesting. If I don't, that's a waste of space and a waste of my time. Algorithmic feeds serve to improve the results such that I am more likely to watch a video that is put in one of those 8 spots compared to if I was just shown trending videos or videos from my subscription box. It is not hostile to me because it saves me time in finding something to watch. You are not being turned into a baby just because you no longer have to search for stuff manually. There is a difference between wanting to find something specific and just wanting to find something interesting. >I've never encountered a situation where an algorithmic feed made things better. One of the best algorithmic feeds is TikTok. After that YouTube's is pretty good to. I suggest trying one of those sites out to just see how effective they are in finding content that interests you. I have found algorithmic feeds have at least made things net better. There is too much information in the world to have to search through all of it ourselves. These algorithms solve the problem of dealing with this large scale of data.

epicide 5y ago
> Algorithmic feeds serve to improve the results such that I am more likely to watch a video that is put in one of those 8 spots compared to if I was just shown trending videos or videos from my subscription box. Right, that's the promise from algorithmic feeds. Sometimes, for some time, it's even the reality! I think there's some middle ground and nuance to be had here, though. I agree that just because there's "an algorithm" doesn't mean it's bad or a worse user experience. Finding interesting videos is a great example. However, I think it's fair to also say that video recommendation algorithms have gotten out of hand. We have a lot of content creators who are influenced in how to make their videos by what is popular. After all, if I make a video that is obviously more successful than another one, that means my audience likes it more, right? What did that video do differently that made it so much more successful? After a long enough time, you end up inadvertently (or intentionally) catering to the whims of the algorithm and not just your audience. So the issue I have with algorithmic feeds isn't that they are outright a bad idea for users. The problem is that it is not just a feature for users. It starts to shape the content on the platform. This can be used to subtly influence content creators to make better content, but without any idea of what goes into these algorithms, we are left to make wild religious guesses. And we have to assume that these algorithms serve the (financial) needs of the companies behind them and their sponsors, first and foremost. If not now, then gradually, by the same feedback loop as happens to the content creators.
lukevp 5y ago
The problem with this is that your suggestions are based on past viewing history, which isn’t always related to future needs. Examples: on one youtube account we primarily watch different types of media on, we watched a Labrinth music video. Neither my wife or I like labyrinth, but now our music section in YouTube is completely consumed by labyrinth music videos. One or two viewings isn’t indicative of us being labrinth fans. This is what we mean by the AI is dumb and makes things worse for the user. The music video section on that account is basically useless now. The same thing will happen with anything - one time I watched a restoration video for an old lamp I had, and YouTube assumed I now mostly wanted to watch antique restoration videos for the next 6 months.
- zimpenfish 5y ago
  Same issue here - if I watch one video from a channel I'm not subscribed to, my recommendations will be flooded with that and related channels. Even picking "don't recommend this channel" only cuts down the spam, doesn't stop it - which is annoying because, sometimes, I've enjoyed what I've watched and I wouldn't mind the odd recommendation ... just not a dashboard full every single day from then on. Oh and "XYZ viewers also watch... [5 videos by XYZ]" is such an egregious failure that I can't help but laugh every time I see it.
- eloisius 5y ago
  God. You watch one ICP video so that you can laugh at the “fucking magnets, how do they work?” line…

version_five 5y ago

Worse, they incentivize "creators" to post "refrigerator door wisdom" type stuff optimized for likes or clicks or whatever (or outrage-bait, but that seems to be more in non-professional forums).