> A feed is a set of algorithmically generated content provided to the user without explicit input.
Uhhhhhhhhhhh no? Wrong & harmful, out of the gate. Feeds derive from "RSS Feed", which are non algorithmically generated: they are each blog's, person's or site's semi-realtime "feed" of their individual content. They are not customized algorithmic content. Web feeds[1] are feeds of web sites.
There's some ok ideas in this article, but the very beginning is a hard-stop "no", wrong, incorrect. It uses an incorrect definition that dooms an idea (feeds) much older than what the article is scolding.
Feeds are wonderful things that allow users & their user agents to create views however the user wants. What the author is talking about is something different. Polluting the idea of the feed by associating it with this Big Tech walled-garden "personal feed" or what-not does injury to the remedy, hurts the chances of real actual & good feeds being able to counter-act the social-network-monopoly "feed"/ween that big tech has us on. This article is harmful.
If you want to talk about the consumer point of view, of someone consuming feeds, a bunch of RSS feeds when combined together forms a "river"[2]. That, too, is non-algorithmic, is an in-temporal-order view of all the feeds.
The problem with feeds is that feeds are not feeds. If instead of feeds we had feeds, users could apply their own algorithmic agents to help them navigate the feeds they subscribe to.
It's an aggregation/vertical control problem. Because big communicative-capitalism commands the means of content creation, content discovery, & content browsing, all mixed together under the vague meaningless "feed" idea, we cant begin to imagine alternatives. If we separate out the content feed from the consuming agent, decouple these systems, we can start to think of different, more pro-user, healthier ways we might approach the feeds we subscribe to. Having everything totalized under one monopoly prevents innovation & exploration of the problem space.
It's important to have the ideas separated out: different channels of information are available to us (a "feed" in the classical/accurate sense). The different approaches to going through the subscribed channels ("feeds" in aggregated/consumed/this-article's sense) is where opportunity lies to find a healthy relationship between ignorance & harmful information-addiction. Currently we have no say in that relationship, other than withdrawl.
From this framing, though, there's only a few reasonably moral possibilities. Like, surely the system of "feeds" as described by the author is not good and should be destroyed, right? The only way for users to have control over "feeds" is for them to not get their content from businesses at all.
> You may be asking: don’t feeds broaden our exposure to new ideas? Yes, they do.
I disagree with this. Feeds are carefully crafted to create feedback loops. Did you click X? You're going to see more X on your feed; showing not-X would make no sense from an engagement perspective. They are echo-chambers by design.
The feed is harmful, but so is sugar. And just like it's saccharine cousin, I don't see it going away any time soon.
Feeds could be crafted to expose us to new ideas and to find the content we are interested in. But instead they are optimized to create an addictive experience and generate revenue.
Twitter removing the linear feed reduced the value of the website for me because I couldn't see all the tweets of the people I follow anymore. I don't want to have to go to each persons twitter page just to see if there are updates.
I think it comes down to the 'free' advertising based model of these companies. As long as any paid advertising is allowed online we can't really trust feeds controlled by other people.
warning: anecdotal. back when i wasn't banned on twitter, it was doing a good job surfacing outside, surprising content, seemingly without repetitous design. i got interesting tech tidbits from surprising places, regularly.
sugar is just one substance. i'm on mastodon & chilling elsewhere. these days there's a little less animosity & hostility in the algorithm vs timeline view, but for a while popular outrage against the algorithm was high. and i didn't buy in. i saw the algorithm as giving me good content, outside content, different things i wouldn't have seen.
i think there are dangerous avenues too. i'd like to know more about how many other people got recommended the "niche" "interesting" content i saw, get some idea how broadscale versus personal it was. but it was working for me. it wasn't an echo chamber. it was surfacing interesting tech topics.
Yes and no. There's no doubt they often act like echo-chambers. However, there are also innocent examples of them enabling discovery, e.g being recommended a knife skills video after watching many cooking videos - it's in the general domain you're interested in, but was probably something you hadn't considered before.
The level of engineering put into creating feeds and distracting you from your purpose on some websites is terrifying. As part of my attempts to lessen my digital addiction, I wrote filters in ublock origin to block off anything on youtube that was distracting (auto suggestions, feeds). I had to target at least 4 different places to get the site looking clean.
There's the home page, the video suggestions at the side of a video you are watching, the video suggestions that pop up when a video ends, and much more. All equally distracting algorithmically-curated rakes that you have to step over. Here are some of the filter strings I used:
Fortunately, ublock origin does most of the targeting for me, although it focuses on ads not content. I'm pretty sure those people work night and day to keep the filters accurate. As for my filters, I've set them over the last couple years and haven't needed to change them since. Changing them probably corresponds to UI overhauls of the site. If it is too difficult or costly for me to filter, I usually just block the site outright.
That's good insight, and mostly aligns with my experience. In the past it was mostly js players and similar api calls via scraping/injection, which are more prone to changing their selectors and interfaces.
(for reference, this was as a FOSS maintainer for a project called BeardedSpice which tapped into a bunch of different media sites for background play/pause/next control).
The theme I get from this post is that technology is becoming more and more frictionless at the expense of our agency.
And when we lose control of what we do next we lose track of where we want to get to.
"The feed" does not fully capture this idea. There are feeds that do not take away your control.
My hobby project is a "feed". But the contents of the feed is fully determined by what items you explicitly upvoted and downvoted.
When you upvote an item, you get stronger connected to other users who upvoted it. Their other upvoted items start ranking higher for you.
When you downvote something, your connection to others who upvoted it becomes weaker. So their future upvote have less weight.
This algorithm is transparent and predictable. This makes it possible to meaningfully interact with it. What you vote on has direct consequences for what you will see in the future. It makes you think about your future self. Where you want to be. When you consider to upvote something you need to answer the question: do you want to get more content from people who found this useful?
If this sounds interesting to you, give it a try at https://linklonk.com/register with invitation code "hn" (a temporary account is created, you don't need to give your email to try). It is very early days, I would appreciate any feedback.
I am interested in what novel ways they'll avoid bubbling - I'd love this network if they could avoid the same bubble you get by choosing your Facebook friends.
Yes, to address this I have an FAQ entry for it (https://linklonk.com/about):
"
Is it a filter bubble?
On LinkLonk you pay attention to those who you chose to pay attention to. In a sense, LinkLonk is a filter bubble.
A filter bubble is a problem when a system chooses content to show to you without giving you clear control or an explanation of how it came up with these recommendations.
On LinkLonk the ranking mechanism is transparent and is easy to understand. LinkLonk does not try to guess what you would like. What you see is controlled by your explicit ratings. For example, when you see a recommendation from users, LinkLonk explains what links you have in common with these users.
"
Also, I think "echo chamber" (compared to "filter bubble") is often used to describe a dynamic that emerges in groups - when members of a group reinforce who is in and who is out of the group. I think LinkLonk avoids this problem by not having the concept of a group. Every user decides for themselves who they want to hear from. There is no boundary to reinforce - no echo chamber walls to erect.
I don't know for sure whether it would become a harmful echo chamber or a useful tool that helps you find high signal-to-noise information. I'd like to give it a try to find out.
I find it interesting that the article has been flagged. It looks like the HN community is shifting towards using flagging, instead of just plain downvoting, as a standrad means of expressing disagreement.
OP here, I'm surprised too. Though this article has since been unflagged, I'm curious what community guidelines people thought were being violated? I'd love to hear any insight from voters into why it was flagged - a semantic disagreement about the word "feed" isn't a justifiable reason imo.
26 comments
[ 3.2 ms ] story [ 64.3 ms ] threadUhhhhhhhhhhh no? Wrong & harmful, out of the gate. Feeds derive from "RSS Feed", which are non algorithmically generated: they are each blog's, person's or site's semi-realtime "feed" of their individual content. They are not customized algorithmic content. Web feeds[1] are feeds of web sites.
There's some ok ideas in this article, but the very beginning is a hard-stop "no", wrong, incorrect. It uses an incorrect definition that dooms an idea (feeds) much older than what the article is scolding.
Feeds are wonderful things that allow users & their user agents to create views however the user wants. What the author is talking about is something different. Polluting the idea of the feed by associating it with this Big Tech walled-garden "personal feed" or what-not does injury to the remedy, hurts the chances of real actual & good feeds being able to counter-act the social-network-monopoly "feed"/ween that big tech has us on. This article is harmful.
If you want to talk about the consumer point of view, of someone consuming feeds, a bunch of RSS feeds when combined together forms a "river"[2]. That, too, is non-algorithmic, is an in-temporal-order view of all the feeds.
[1] https://en.wikipedia.org/wiki/Web_feed
[2] https://github.com/scripting/river5
Let's talk about the concepts instead!
The main problem is that users don't have control over their feeds anymore. Which business models could change that?
It's an aggregation/vertical control problem. Because big communicative-capitalism commands the means of content creation, content discovery, & content browsing, all mixed together under the vague meaningless "feed" idea, we cant begin to imagine alternatives. If we separate out the content feed from the consuming agent, decouple these systems, we can start to think of different, more pro-user, healthier ways we might approach the feeds we subscribe to. Having everything totalized under one monopoly prevents innovation & exploration of the problem space.
It's important to have the ideas separated out: different channels of information are available to us (a "feed" in the classical/accurate sense). The different approaches to going through the subscribed channels ("feeds" in aggregated/consumed/this-article's sense) is where opportunity lies to find a healthy relationship between ignorance & harmful information-addiction. Currently we have no say in that relationship, other than withdrawl.
I disagree with this. Feeds are carefully crafted to create feedback loops. Did you click X? You're going to see more X on your feed; showing not-X would make no sense from an engagement perspective. They are echo-chambers by design.
The feed is harmful, but so is sugar. And just like it's saccharine cousin, I don't see it going away any time soon.
Twitter removing the linear feed reduced the value of the website for me because I couldn't see all the tweets of the people I follow anymore. I don't want to have to go to each persons twitter page just to see if there are updates.
I think it comes down to the 'free' advertising based model of these companies. As long as any paid advertising is allowed online we can't really trust feeds controlled by other people.
Do you have tips on if / how you manage feeds, if you do?
I'm trying to figure out a twice a year sort of audit to do for myself. I haven't figured it out yet.
sugar is just one substance. i'm on mastodon & chilling elsewhere. these days there's a little less animosity & hostility in the algorithm vs timeline view, but for a while popular outrage against the algorithm was high. and i didn't buy in. i saw the algorithm as giving me good content, outside content, different things i wouldn't have seen.
i think there are dangerous avenues too. i'd like to know more about how many other people got recommended the "niche" "interesting" content i saw, get some idea how broadscale versus personal it was. but it was working for me. it wasn't an echo chamber. it was surfacing interesting tech topics.
There's the home page, the video suggestions at the side of a video you are watching, the video suggestions that pop up when a video ends, and much more. All equally distracting algorithmically-curated rakes that you have to step over. Here are some of the filter strings I used:
Unreal, in my opinion, how firm of a hold these kinds of sites can get over your brain through dopamine addiction and distraction.I'm curious, how often do you have to update the selectors?
context: I've maintained various website parsers over the years, and it seems to be a constant struggle, depending on the targeted function.
That's good insight, and mostly aligns with my experience. In the past it was mostly js players and similar api calls via scraping/injection, which are more prone to changing their selectors and interfaces.
(for reference, this was as a FOSS maintainer for a project called BeardedSpice which tapped into a bunch of different media sites for background play/pause/next control).
This is like "assume a spherical cow" or "define red to mean green".
> "feed", as used colloquially definitely refers to what OP means. So, by giving a definition OP makes clear which meaning we're talking about.
Feed is an overloaded term, so this was the best way I could disambiguate them. Any suggestions for alternate names that are more clear?
And when we lose control of what we do next we lose track of where we want to get to.
"The feed" does not fully capture this idea. There are feeds that do not take away your control.
My hobby project is a "feed". But the contents of the feed is fully determined by what items you explicitly upvoted and downvoted.
When you upvote an item, you get stronger connected to other users who upvoted it. Their other upvoted items start ranking higher for you.
When you downvote something, your connection to others who upvoted it becomes weaker. So their future upvote have less weight.
This algorithm is transparent and predictable. This makes it possible to meaningfully interact with it. What you vote on has direct consequences for what you will see in the future. It makes you think about your future self. Where you want to be. When you consider to upvote something you need to answer the question: do you want to get more content from people who found this useful?
If this sounds interesting to you, give it a try at https://linklonk.com/register with invitation code "hn" (a temporary account is created, you don't need to give your email to try). It is very early days, I would appreciate any feedback.
A filter bubble is a problem when a system chooses content to show to you without giving you clear control or an explanation of how it came up with these recommendations.
On LinkLonk the ranking mechanism is transparent and is easy to understand. LinkLonk does not try to guess what you would like. What you see is controlled by your explicit ratings. For example, when you see a recommendation from users, LinkLonk explains what links you have in common with these users. "
Also, I think "echo chamber" (compared to "filter bubble") is often used to describe a dynamic that emerges in groups - when members of a group reinforce who is in and who is out of the group. I think LinkLonk avoids this problem by not having the concept of a group. Every user decides for themselves who they want to hear from. There is no boundary to reinforce - no echo chamber walls to erect.
I don't know for sure whether it would become a harmful echo chamber or a useful tool that helps you find high signal-to-noise information. I'd like to give it a try to find out.