People filter themselves into bubbles. That's how websites like Twitter and other social media work. You choose who to follow. You choose who to block. Getting rid of bubbles would be a...
stop filtering people into bubbles
People filter themselves into bubbles. That's how websites like Twitter and other social media work. You choose who to follow. You choose who to block. Getting rid of bubbles would be a fundamental restructuring of how most social media works.
Now, I will agree you can be more filtery/bubbly or less depending on how you do things. Algorithmic front page vs chronological is the most obvious example.
The thing, is, though, under the current model, the two are effectively the same. I mean, one requires initiative from the user, and one is just an algo working in the background, but they...
The thing, is, though, under the current model, the two are effectively the same. I mean, one requires initiative from the user, and one is just an algo working in the background, but they ultimately tend toward the same result: people are fed, or feed themselves the garbage they're already into.
Algorithmic bubbles can be different. Like, the joke is that the Youtube algorithm is "would you like to become a nazi today?" but it could just as well be the opposite. You could have a "tend towards the middle" algo, for example. Or a "try something completely new but still tangential to your interests" algorithm.
The bubbles people make for themselves, though, tend to all be the same, and are arguably just as dangerous. I think it's weird to say "right companies should 100% stop suggesting content," and at the same time say "It's important to always give individuals all the choices and let them figure things out for themselves," because people tend to make garbage choices. (I'm not excluding myself from this, to be clear.)
Idk what the answer is, but I think suggestions and bubbles are still useful. We should figure out how to benefit from them, while also avoiding the dangers, and making sure we don't crush human autonomy somewhere along the way.
Reliance on algorithms is fundamental to the value proposition that these companies are making. The whole point is to take (expensive) human judgment out of the equation and replace it with...
Reliance on algorithms is fundamental to the value proposition that these companies are making. The whole point is to take (expensive) human judgment out of the equation and replace it with (cheap) algorithms.
I agree that it's never been effective - in my own experience, the only thing that has worked is that people self-select into communities that have human moderation staff willing to apply sound...
I agree that it's never been effective - in my own experience, the only thing that has worked is that people self-select into communities that have human moderation staff willing to apply sound judgment. The point I wanted to make is that these companies will never respond by providing that because it is fundamentally in opposition to their business model.
From a dollars and cents perspective, the average revenue per user (ARPU) on these sorts of sites is $5-$10 (scale that up to $20-$30 if we talk just users in developed Western countries). Your phone company, cable company, or bank probably have an ARPU higher than that per month - how much human service, and in particular, sound exercise of human judgement, do you get from them?
This isn't meant to dismiss your concerns or arguing people aren't being harmed. I agree that they are and I think the sites in question have a moral obligation to do something. That said, I don't expect that they will, or if they do, the response will be the most easily automated, hamfisted response possible (i.e. if someone is being harassed, ban the victim from their platform thereby "stopping" the abuse).
People filter themselves into bubbles. That's how websites like Twitter and other social media work. You choose who to follow. You choose who to block. Getting rid of bubbles would be a fundamental restructuring of how most social media works.
Now, I will agree you can be more filtery/bubbly or less depending on how you do things. Algorithmic front page vs chronological is the most obvious example.
Besides that, I largely agree.
The thing, is, though, under the current model, the two are effectively the same. I mean, one requires initiative from the user, and one is just an algo working in the background, but they ultimately tend toward the same result: people are fed, or feed themselves the garbage they're already into.
Algorithmic bubbles can be different. Like, the joke is that the Youtube algorithm is "would you like to become a nazi today?" but it could just as well be the opposite. You could have a "tend towards the middle" algo, for example. Or a "try something completely new but still tangential to your interests" algorithm.
The bubbles people make for themselves, though, tend to all be the same, and are arguably just as dangerous. I think it's weird to say "right companies should 100% stop suggesting content," and at the same time say "It's important to always give individuals all the choices and let them figure things out for themselves," because people tend to make garbage choices. (I'm not excluding myself from this, to be clear.)
Idk what the answer is, but I think suggestions and bubbles are still useful. We should figure out how to benefit from them, while also avoiding the dangers, and making sure we don't crush human autonomy somewhere along the way.
I know that's vague.
Reliance on algorithms is fundamental to the value proposition that these companies are making. The whole point is to take (expensive) human judgment out of the equation and replace it with (cheap) algorithms.
I agree that it's never been effective - in my own experience, the only thing that has worked is that people self-select into communities that have human moderation staff willing to apply sound judgment. The point I wanted to make is that these companies will never respond by providing that because it is fundamentally in opposition to their business model.
From a dollars and cents perspective, the average revenue per user (ARPU) on these sorts of sites is $5-$10 (scale that up to $20-$30 if we talk just users in developed Western countries). Your phone company, cable company, or bank probably have an ARPU higher than that per month - how much human service, and in particular, sound exercise of human judgement, do you get from them?
This isn't meant to dismiss your concerns or arguing people aren't being harmed. I agree that they are and I think the sites in question have a moral obligation to do something. That said, I don't expect that they will, or if they do, the response will be the most easily automated, hamfisted response possible (i.e. if someone is being harassed, ban the victim from their platform thereby "stopping" the abuse).
Maybe something will come out of this that will help all online communities.
Here is a similar news article, focusing on a few different details.