Algorithms don't cause filter bubbles, people do
news.findka.com
news.findka.com
I did the same thing with reddit. Unsubscribed to subreddits I didn't want to see anymore and subscribed to ones I did, and now my reddit front page is 100% programming links, gamedev links, 3d modelling, cooking, and a few other niche interests I have.
I don't think it'd even be all that hard. What you'd want to do is have a mode for contentious topics whereby personal identities are hidden, so there's no personal investment and only ideas are presented.
And you'd want to be able to cite problems with a claim, i.e. you select some text to flag "this is confusing / inflammatory / unsupported," so that people don't go in circles because they don't understand each other or can deal with something that's insulting by having the person tone it down.
I think that applies to the Internet in general, including HN.
I've never seen an online community improved by allowing more object-level political discussion.
This is purely my own experience, but I find that I need to be in the right mindset to read something that's going to challenge me. If I'm scrolling for a few hits of updates from people I care about, I'm not going to be in the right state of mind to read something that's going to get me riled up. The bubble may or may not be inevitable, but it's a necessity. Being aware of feedback loops and managing your own media intake is probably more important than finding ways to prevent those loops entirely, IMO.
The obvious result of this was that I realized nothing changed when I did this, except for less useless info entering my eyes, and I gradually stopped using FB almost entirely. Now I check if there are any interesting events on every week or two, and that's it (although I still use messenger - attempts to get friends and family to switch have had mixed results).
- To prevent identity theft by someone creating an account in my name.
- To access some local event planning that only happens on Facebook, e.g. local motorsport competition/HPDE events
- To access Facebook Marketplace
Everyone does. There's nothing wrong with it. People at the coffee shops or the pubs were often people with similar interests.
The problem is that we exited the culture wars in around 2009. We went into societal crisis. For the next 10-15 years or so we will be looking for the collapse of the crisis.
The funny thing is that people still think it's the culture wars. Especially the mass media. It's not. It's why culture wars like Gamergate or Greta Thunburg are not winning anything at all for their cause.
The reason people are very concerned about these filter bubbles and the problems of social media is because those were fine during culture wars. They are huge problems now we are in the crisis era.
The problem? The collapse of the crisis is always violent revolution or a world war that does not hold back.
People want to break these bubbles in order to prevent what is rather inevitable.
The funny thing is that nobody can predict who or what is the revolution. Which side has to lose? There's a bit of a problem. The democrats want to ban guns and generally dislike guns as a group.
So if the revolution happens, who loses is pretty clear.
It's a problem because it alienates you from your irl community
Im in an online bubble because im talking about a niche idea?
What if I start talking about Django ModelAdmin templates? That would be quite niche. I bet there's only 5 other people in my city who could have that conversation with me.
Point Im making is that yes it's quite niche and quite the bubble. I never denied the bubbles exist. I explicitly said everyone lives in a bubble and there's nothing wrong with it.
The important reality is the rest of the subject. We aren't in culture wars anymore. We are in crisis akin to https://en.wikipedia.org/wiki/Glorious_Revolution 77 years later https://en.wikipedia.org/wiki/American_Revolution 78 years later https://en.wikipedia.org/wiki/American_Civil_War 78 years later https://en.wikipedia.org/wiki/World_War_II
2021 - 1945 = 76
>It's a problem because it alienates you from your irl community
what IRL community? nobody is allowed to see anyone else. It's also not like this is only subject I know about. Im not getting alienated.
Unfollowing and tailoring content to your interest is great and makes a lot of sense. The issue arise when dealing with news. If Facebook is your main news source and you decide to unfollow any medias, you can end up in a bubble where the world news you are presented with are really skewed to reinforce your world view which ultimately could lead you to full own fake-news consumption.
That being said, inflammatory political posts by your friends are not news.
Sure, I decide who to interact with. But I make that decision based on how they post. They post, in turn, based on how other people react to their posts.
About half of the things on my FB feed never get a single like or comment. I'm not sure why I still post them. Some of my misses:
"Judge permanently dismisses charges against [Breonna Taylor's boyfriend]": https://www.cbsnews.com/news/kenneth-walker-breonna-taylor-j...
"TikTok assures users worried about Myanmar that they're aggressively monetizing the situation": https://www.theonion.com/tiktok-assures-users-worried-about-...
"Here's why Texas is still without power ... and why other states are vulnerable too": https://yaleclimateconnections.org/2021/02/why-the-power-is-...
Here are some things that got multiple interactions:
this image mocking TX Gov Greg Abbott: https://i.redd.it/wlq9vk0k9ri61.png
this meme about Gamestop: https://i.redd.it/vdfzutt6b4e61.jpg
It's easier for me to post low-info, high-emotion, easily-consumed polit-ideological junk food and get likes than it is to get attention with a link containing some information or argument, or even just inoffensive satire. Explanations are out, newsfeed-optimal screenshots of brief text are in.
It's very misleading to ignore the influence of the environment when considering how people choose whom to interact with.
yah, that’s a form of the fundamental attribution error, a pretty common error in any discussion. we won’t get more civil discussions anywhere without systemic changes that lure us further into such errors by design.
Just like a garden, for a social media feed to grow it needs to be taken care of and curated. There's a block and mute button for that.
I've tried to create a bubble on Instagram of the content I want to see. But Instagram is more agressive about pushing algorithmic content at me against my will. After days of trying to train the algorithm, I got it to only show me real photos of travel and nature. But slowly new things are always creeping in. Liking one photo can set off echoes that are hard to get rid of. If I accidentally like a photo with a pretty girl in it, or someone's sweet workstation setup, or a map, or a meme, Instagram starts filling my feed with low quality junk related to those topics. Even the travel and nature photos are low quality junk. The inability to set up a feed of content I actually want to see (friends and genuine non-influencer travel photos with under 200 likes) is pushing me away from Instagram as a platform.
Facebook's news feed is so bad that I never scroll it at all.
For me, RSS via Inoreader is my new way to curate most of my content. The only problem is that most of my friends don't post RSS friendly content. Including friends in filter bubbles seems to be an area that no existing platform does a good job of anymore.
> There is always an algorithm. The default one promotes rage, controversy and misinformation. But we don’t have to stick with the default.
And I guess the real problem is that the financial incentives are set up such that sites like Facebook are motivated to choose[1] algorithms that are most suited at stoking engagement via rage and controversy.
[1] https://www.wsj.com/articles/facebook-knows-it-encourages-di...
There's also a lot of possible variation in the strength of the bubble. Suppose I click on "Fox News". Should FB feed me more right wing news? Probably: as you say, people want news they are interested in, perhaps for good reasons. Should it try feeding me more extreme stuff, to see if I'm interested in QAnon? Maybe not, but it's easy to imagine how recommender algorithms could end up doing that.
This is a low-effort content marketing fluff piece that just says this company's recommendation engine has a "serendipity knob." It avoids the pernicious bits like how business models built on eyeball engagement result in a death march towards maximizing engagement metrics, and engineering the most addictive filter bubble has proven a pretty effective way to juice some PM's quarterly numbers.
Make a post on how to break that model, don't say "we have a serendipity knob tuned to 9.4%!"
Same question for a fast-food wrapper. Do we think McDonalds should serve its food loose?
A restaurant will wash their stuff to make sure that you're a return customer. But a big advantage of once-opened / once-disposed containers is that I can suddenly trust random street vendors with no relationship to me.
This type of fraud/malpractice is much more likely to be seen in small restaurants with little oversight and loose processes that at a multi-million dollar bottling plant.
Its Street Vendor #455. Who is refilling reusable bottles and selling them to his customers. I'm talking about people sitting with coolers and ice, selling $1 soda or $1 water.
I think you’re looking for a problem that doesn’t exist in the real world.
Or we could have laws and inspectors for that at the bottling facilities because the above is not going to work.
The same assurances I have that they didn't let a rat piss in it at the factory the first time around; inspections and safety regulations.
I get my milk in reusable glass bottles. I have absolutely zero concern that they forgot to wash them between uses.
That's the thing. A disposable container guarantees that the bottle came from Coca Cola.
---------------
> I get my milk in reusable glass bottles.
But you don't get your reusable milk from a street vendor. You trust your deliveryman, you trust the company they work for. You have a working relationship with them. (Or maybe you pick it up from a grocery store that you trust).
I'm talking about a street vendor: someone you have absolutely no relationship with. How do you know that you can trust the street vendor?
A sealed disposable bottle offers guarantees that reusable containers do not. That's simply a fact.
My reusable milk bottles come with a small plastic top that's recyclable and functions as an anti-tamper seal. No reason you couldn't do the same with soda; plastic Coke bottles already have this.
Unless you buy a sealed bottle to begin with, you don't really know if it was reused. I admit that you have a good point: if the glass bottle had an appropriate one-time seal on it, that would be proof enough for me.
Really, if you are so concerned about random street vendors, you should consider that it is not hard at all to open a plastic soda bottle without breaking the seal ring on the cap, refill the bottle, and put the cap back on.
It's hard to win. Thin plastic = less materials consumed and less fuel spent on transport. Switching to thick plastic might make it harder to litter, but the flak will just shift to how wasteful the packaging is.
2 is because parts fall off of people's cars. 3 is because construction material will fall out of the pickup truck beds that are used by construction workers.
But the food packaging waste is dropped by people. If we accept the UX principle that it's easier to change the system than human behavior, then the best way to reduce the volume of litter would be to reduce food packaging waste. For example, if fast food, like McDonalds, were packaged in rapidly bio-degradable and low-pollutant things, like simple, un-dyed paper wrappings, then the amount of waste in the environment would be greatly reduced.
But then...
> newsletter, selected based on your interests
Maybe I misunderstand what a "filter bubble" is? Because this seems like a type of filter bubble to me. Perhaps a little looser (and less rage-inducing) than Facebook but it still seems designed to keep people reading content only within their bubble-of-interest while filtering out the rest.
What I'd like to see is something that can select for quality, and not for interest.
I think this is what you often get with human manual curation. For example, I can happily read a well written well researched thought piece even if the content speaks to opinions I don't currently hold. Or news source that at least appear to have done their research and provide ample evidence and facts, no matter what the content is for.
Similarly with entertainment too, friends can recommend some new great war movie even if I'm not generally into war movies, or some new great book even if I rarely read fiction, etc.
I don't really know how to judge quality though. When someone asks me what kind of music I like, I tend to answer I just like good music. What does that mean? I'm not sure. Could machine learning figure out what that means if it trained on my personal liked music tracks? Would it learn the qualities I'm interested in, or would it end up only learning my personal preferences at the time?
> That, however, is one of the reasons I spend so much time working on recommender systems. There is always an algorithm. The default one promotes rage, controversy and misinformation. But we don’t have to stick with the default.
It means that people would create their own filter bubbles, even in the absence of an algorithm, so blaming algorithms or trying to do away with them if your goal is to eliminate filter bubbles is moot.
They were granted to us by the Ixians.
This is one of those annoying articles where someone talks about their specific interest as if it's a universal truth for an entire domain of problems. For an Amazon product listing or a Buzzfeed news page that thrives on users seeing new and exciting stuff to click on and buy, sure, the recommender system needs to avoid getting stuck on one specific type of content. For your Facebook newsfeed or YouTube video sidebar where divisive politics rules the day it is not absolutely not goal because that would stray in to challenging user's beliefs, and that would push people away.
I have relatives who do not get a print paper, but tend to read the The New York Post and The Washington Examiner on-line. The watch Fox News and at some point probably listened to Rush Limbaugh.
Do we get rather different views of what is going on? I can't see how we could not. Did an algorithm have anything to do with this? You would have to dramatically stretch the sense of "algorithm" to say so.
Shameless plug: https://chrome.google.com/webstore/detail/filterbubble/amgno...
This line would be a full-stop for my librarian SO:
> It’s referred to in the literature as “serendipity.”
But otherwise I'm very grateful for this article.
For it reminds me that someone should update (adapt) the Propaganda Model for social media. https://en.wikipedia.org/wiki/Propaganda_model#Filters
That would be far more constructive than the alternative theories floated, eg Eli Pariser (filter bubble), Ben Thompson (aggregation model), and this technopositivistic missive.
Has anyone asked Chomsky? I've looked, but no joy.
--
Edit:
> There is always an algorithm. The default one promotes rage, controversy and misinformation. But we don’t have to stick with the default.
Clap louder, and surely Tinkerbell will fly.
Of course it's a goal. They're just not very good at achieving that goal. I wouldn't expect someone who makes their living building these things to be unbiased on the question of whether or not they're currently a bad thing.
If I search incognito youtube for "Trump Speech" and select the first video, 3 of the next 4 "up next" are also Trump speeches, and the other one is "McCarthy SHREDS Pelosi". If I click on of those 4, the next "up next" videos are different, but also 3 of the first 4 are Trump speeches. The other one is a Trump-Biden debate. This is a filter bubble.
Sure, human psychology plays a part, that's why algorithms like this work in the first place. But compare this to the opposite extreme pre 24-hour news networks and there was simply much less opportunity for playing into this. That's what algorithms provide: The ability to "instantiate" filter bubbles much more readily than content that is passively delivered without consideration for past choices, and the consumer has to make all of their choices without that outside influence.
[1]: https://slatestarcodex.com/2014/12/17/the-toxoplasma-of-rage...
"There is always an algorithm. The default one promotes rage, controversy and misinformation"
This guy is a massive idiot and he doesn't know it yet.
https://web.archive.org/web/20210309163058/https://news.find...
https://www.algotransparency.org/
https://arxiv.org/pdf/2003.03318.pdf
https://www.brookings.edu/techstream/how-youtube-helps-form-...
No we don't. But does anyone think that Twitter would have risen as it has if it did not select for rage and controversy? There is an algorithm on another layer of the stack that filters out more neutral content: The free choices of users in an open market. If Twitter (et al.) did make an effort to screen out its most viral content it would simply lose market share to a platform that doesn't.