▲ 88 points
back
135 comments
By 'architecture of the internet', the authors mean the nature of social media feeds.
The problem is more specifically "algorithmic feeds" which isn't require by or exclusive to social media. For example, news sites and media sites like Youtube and Spotify (which arguably have social aspects, but most people don't use them like social media) also contribute in similar ways. The root problem is the algorithm optimizing for attention mixing with human nature that tends to make negative reactions more powerful than positive reactions which causes the algorithms to create a sort of polarization death spiral.
If we are going to be pedantic, let's do it correctly. The specific, direct problem is advertising, where the money flows. Algorithmic feeds are designed to capture attention, and the reason for attention capture is eyes on advertisements.
Having identified the true root of the problem, I would recommend directing resources towards dismantling advertising. Focusing on anything else is wasted effort.
And, as usual, "risk/threat to democracy" is used to mean "support for parties I don't like"
It wasn't long ago that the Twitter shoe was on the other foot, and many of those complaining now were quite happy to endorse the right of private companies to promote/suppress speech at will (with no hint of irony regarding their alleged ideological views on private companies)
What I don’t get about these centrist takes is that even if you believe this, can you not recognize the difference in motivation and intention of the speech? For example, to me it’s wild to equate something like Democrats wanting to promote Covid safety and Republicans wanting to promote the idea that Californian elections are fraudulent. One is an earnest attempt for public good and the other is a cynical attempt to undermine democracy.
Much agree on this. The article is clearly biased. But one thing it mentions is that even when disabled people don't change their mind. Perhaps some politicals views are more rational and sound. Can the reverse be true? We don't know.
>And, as usual, "risk/threat to democracy" is used to mean "support for parties I don't like"
It can also mean highly influential support for ideologies I don't like - like fascism, authoritarianism and ultranationalism.
> many of those complaining now were quite happy to endorse the right of private companies to promote/suppress speech at will
A few years ago, bog-standard content moderation was limiting the reach of enormously+reasonably unpopular ideologies like fascism, authoritarianism and ultranationalism. Groups who were profoundly unhappy with these limits would bullhorn complaints of intentional suppression and censorship.
With the release of the Twitter Files (which exposed content moderation), it became clear that many folks were unable to differentiate between actual, long-established content moderation methods and actual directed suppression and censorship.¹
This deep misinterpretation seemed to flow from the ignorance of what content moderation looks like at scale. That core misunderstanding was often amplified and made worse when historically-moderated individuals filled in that vacuum with their long standing preconceptions.
The upshot are today's efforts to raise the visibility of far-right viewpoints thru coordinated crafted messaging² and thru actual suppression of non-right viewpoints thru new controls over platforms and thru often unaccountable misuse of governmental powers.
Our present conditions seem to well reflect and align with the article author's analysis.
This might be a "yes, and" moment, but the article specifically points the largest feeders of contemporary thought leaning algorithmically in one direction. It doesn't matter if that direction is left or right, either way shaping thought in one direction is anti-democratic and that's what should worry all of us.
I think there’s too much focus on the internet and social media here. We should look back to the printing press as the origin and mass media, and trace the development through to radio and television. The risk for democracy is not social media per se, but mass media.
It goes beyond that. Even chat platforms can be a problem now. IMO, I'm no sociologist but I'd love the viewpoint of one, human societies were very much non flat in terms of information, and cheap infinite internet collapsed the thin hierarchical nature of information-sharing and communication.
If that's what they mean, fully agreed.
There's no clean separation between those things. The weakness and inadequacy of HTTP(S) and other protocols actively funnels people into the centralised services of big providers. It creates a world where storage is brittle and content is ephemeral, both directly due to its own failings and because it pushes people towards big providers who increasingly like things that way; and so on. Now human nature would be enough to tend to draw a lot of people towards lowest-common-denominator options, but a system which makes the alternatives frictionful and downright painful doesn't help either.
Isn’t the title inappropriate then? Shouldn’t it include “social media”?
Thanks, saved me a read.
And the feeds are largely the way they are due to unregulated greed.
And by democracy they mean... people agreeing with each other and voting for the correct parties:
broadband reduced civic participation, eroded social trust, and boosted voting for extreme-right and populist parties in Italy and Germany.
Is the "extreme-right" party in Germany still chaired by a brown lesbian woman?
I couldn't read this article because Science.org left Bot Fight Mode or Super Bot Fight Mode enabled in their Cloudflare settings, causing me to be blocked by a “security verification”. If you use Cloudflare, disable bot stop modes by going to dash.cloudflare.com and selecting your domain and then clicking on “Security” and then clicking on “Settings” and then using the buttons to disable Bot Fight Mode or Super Bot Fight Mode.
My experience with dead internet theory isn't that bots are making the internet unusable, it's that anti-bot measures are making the internet unusable. It's like antivirus software, all over again, except this time I don't get choose whether or not to run it.
I was going to give an archive.is link, but they're blocked too.
Do you know what sort of browser configuration triggers the Bot Flight Mode detection?
I'm not using anything too esoteric (Firefox Developer Edition, highly tweaked + extensions).
We learned that computer security had issues during Viet Nam. The problem was studied, lessons were learned, and the problem was actually solved, about 50 years ago. Unfortunately, the PC computer wave washed interest in those solutions down the drain.
Because of this, Windows, MacOS, Linux, are all insecure by design.
Because of that, you can't really host your own servers on the internet, unless you're a die-hard IT guy.
Because of that, the only game in town is walled gardens.
Because of that, everything has to be paid for, thus the algorithm, thus the rot of society.
All because we thought Unix's security model was good enough.
I think it's the architecture of the web, specifically, the fact that you can't use HTML to mark up hypertext, that is the root of it (or maybe my other thread about OS structure was right... you decide)
All this morning, a thread has been building in my mind. On Twitter, you read something, and a witty come-back denying yet affirming part of the original post, and feel tempted, strongly tempted, to try to add to that cascade. It's the structure of Twitter, really.
Here on HN, we try to do the same, but at least we're working with higher quality resources. The problem is that you can't weave a thread together without having to expend thousands of words to replace what should be something much simpler, using a few transclusions, to provide examples of what you mean, something much more information dense, and far less verbose.
Thus, it's much easier to just try to demolish something, by pointing out its most visible flaw, in a few sentences, rather that to build up a novel observation, by crudely trying to provide context with thousands of words.
> Those algorithmic biases have demonstrable behavioral consequences.
The algos optimize for engagement, which can roughly translate into the people drive the algos, as they would stop watching or visiting or commenting, if it was not something they wanted to engage in.
So in some ways, is this not democracy to the max?
I wonder if articles like these don’t like the outcomes, or the reflection of society that the algos create. And thus attack them, because they would rather curate and limit conversation and expressions on the internet they don’t like or agree with.
The danger to democracy has always been uninformed voters.
Now it is mis-informed voters.
> The algos optimize for engagement
That's what we're told, anyways
It isn't too unreasonable to think about that there might be an invisible thumb on the scales for any of these algorithms
When people say "democracy" these days they really mean something closer to "technocracy". (Often they mean technocracy, laundered through democracy)
I think your supposition is correct. I think there is a common hypocrisy to the person craving democracy while showing revulsion at revealed preference. Many otherwise smart people can't seem to look at society without averting their eyes.
Edit: Grammar.
Democracy is a tricky thing. It’s not as simple as “whatever the majority of people wants, goes”, and has always been recognized as such.
Classical liberalism requires certain rights and protections to every member of society in ways they could be perceived as “anti democratic” if for example a minority group is widely hated.
Generally speaking all of this requires some level of rules and forbearance, and a political “playing field” where disputes can be ironed out.
Part of what is required for this to work is a shared epistemology. This has historically been provided by journalist and academic elites, but it the thing that is being eroded by social media.
The problem is now everyone can choose their own reality, but that reality may just be completely not true. This was a well known phenomenon on the right but it’s happening a lot more on the left as well, with it being taken as a fact that everyone in the US is poor and struggling even though that is not true at all.
The net effect of this is that “charismatic” reactionary parties that are detached from reality perform better, because memeing wins elections better than doing things for constituents. The link was always a bit tenuous but now it’s completely broken and we’re seeing the rise of anti intellectual parties everywhere.
A simple improvement to existing social media would be to allow users to see other users' feeds. Just seeing things from a different perspective.
Why not:
(1) directly fund studies and reproductions of studies (promising ahead of time to publish the results, even if negative) targeting the exact issues they're concerned about
(2) writing and publishing extensively to show people the results and help them arrive at a correct interpretation of the data
(3) make a public commitment ahead of time to change opinion based on what the data says, and not to overstate underdetermined theses
... instead of spending money trying to control the political narrative?
That would simply be science doing science -- which has always threatened the establishment because it's accountable to reality, not authority.
Science rightly done never claims authority, just reports on what the data says. Truth is powerful enough on its own.
It's not a panacea or a magic fix for human nature, but one of the root causes of this is that the underlying architecture of the HTTP(S) Web is just inadequate. The world needs (technically viable and widely-used systems of) content-addressable storage: inherently achivable, mirrorable and recoverable, properly supporting intermittent connections, providing the stability which is the necessary (though not sufficient) base for building things like annotations and back-linking. That certainly can't force people not to choose the laziest and stupid options, but it really can't hurt if at least the underlying technology doesn't make doing anything but the laziest and stupidest thing inherently hard, esoteric and unrewarding. Instead we've created TV on the computer from the visionary Doug Engelbart manifesto Don't Create TV on the Computer. Worse, some people still seem to be trying to pat themselves on the back for the supposed pragmatism and savviness of those decisions, even while at the same time using their other hand to wave a fist at the Big Tech incumbents, content farms and grifters which they gave a structural advantage to. There aren't many things which should be a higher priority, and which are a bigger blocker of general improvement, than the continuing lack of widely adopted and widely adoptable content-addressable storage. Need to do something big about that, folks, and promptly.
Would social media be the bane it is, if:
1. There was no algorithm tweaking your feed. Promoting something and suppressing something else.
2. Creators were not paid.
3. Advertisements were randomly allocated irrespective of the content.
4. There were no such thing as likes.
5. Users had the option to pay for an ad-free experience.
Probably yes. The misperception of general consensus would be skewed by the bubble you follow. For instance, if all the people you follow are flat earthers, you might form the perception this is a sensible point of view. Same for vaccines, the war against Iran, Cuba, and so on.
Any story about threats by the Internet to democracy that revolve around Twitter has to account for the fact that only a minute portion of the electorate ever looks at Twitter.
The article mentions basically all major social media though.
Besides, even if it was just about twitter, it can only take a small portion of the population to swing an election. Word of mouth is also downstream from twitter. People might not see something on twitter, but they might hear it from someone who saw it there.
I couldn't agree more. One day I uninstalled twitter(x) and I just kinda forgot about it. A couple of times I tried to look at where the icon used to be and never really felt the urge to reinstall.
I like to think that I am not alone in this and this happened to hundreds of thousands of people. When you overly optimize for engagement at some point you cause burnout and loss of interest. It felt funny seeing musk claim that all twitter statistics were going up without realizing the cost of it. Social media has to strike a very strong balance to keep you engaged, but not too engaged.
It’s not about social networks but algorithmic feeds. The same issues pop up with YouTube or any other website that shows you content that it thinks is more likely to engage your attention.
For the American ideal of democracy or the European ideal of militant democracy?
So do the limitations (and requirements) of hardware and operating systems. And corporations and billionaires financing and supporting antidemocratic systems and politicians.
Modern smartphones could easily be meshnet nodes, but they don't really support P2P networking.
See: FireChat, Bitchat (removed from the Chinese app store), Airdrop (Apple limited its functionality in China)
Extremely biased study. It assumes voting for right wing parties is naturally bad.
TL; DR: Develop and deploy algorithms that downrank or deprioritize anti-democratic, extremist, or polarizing content.
Just call your opponents anti-democratic, extremist or polarising and here you go. Democracy!
Every single one of these "internet is a threat to our democracy" takes is really about a few things, none of which is a threat to democracy.
1) Hand-wringing about information disintermediation: previously, institutional gatekeepers filtered information and interpreted it for the public. Now, the public sees raw information and forms its own judgements.
2) Social media has cut revenue streams for the sorts of organizations that bleat non-stop about how social media is a thread.
3) Weakening of ability of the institutional class to censor defectors and promulgators of inconvenient facts, which disaffected former censors call "disinformation".
Far from being a threat to "democracy", the internet is the best thing that's ever happened to it. Social media and the internet more broadly have enabled an unprecedented increase in breadth and depth of public participation in the marketplace of ideas. Those who don't like the result never liked democracy.
It's exhausting, this ceaseless cacophony of high-minded bullshit. I'm sick and tired of hearing people exclaim that the internet is a danger to "democracy" when, really, the problem is that the internet produces democratic outcomes they don't like.
> Hand-wringing about information disintermediation: previously, institutional gatekeepers filtered information and interpreted it for the public. Now, the public sees raw information and forms its own judgements.
I mean is the information raw really? How raw was #metoo or would you rather meet a man or a bear in the woods. The internet is super-curated. There’s like super obscure intellectual woke all over x/twitter; The opposite of raw, that’s what people like about it! Raw would be a _substantial_ improvement over the like bizarrely curated shit we have everywhere now.
I really haven’t trusted Nature and Science for about two decades. These are captured entities. The value of a publication in Science or Nature is questionable. This is a consequence of the corporatization of science in academics.
“Nature was, and is, a commercial enterprise, owned by the privately held company Macmillian publishers. . .”
Get off the Internet, go read a book. “Plastic Fantastic: How the Biggest Fraud in Science Shook the Scientific World” - Eugene Samuel Reich.
The bullshit artists are at it again.
These studies of what is suppressed on social media somehow always overlook that Facebook bans all white nationalist content, or the purges of right-wingers from reddit. Censorship the authors agree with does not "create risks for democracy".