130 Comments
Dec 16, 2022Liked by Noah Smith

Surprised to see no mention of Reddit. Reddit is basically what you mention, smaller interest based communities with community moderation.

Of course, Reddit can be pretty toxic though too, lol.

Expand full comment

This is perceptive. An odd part to the story is one company that seemed great at cracking this sort of thing also realized this issue, but its solution completely failed - Google, with Google Plus.

Their big advance was supposed to be "Circles", where you could sort your contacts into different groups whom you'd then post different content to, basically have different social networks with within the Plus umbrella.

There's a lot of reasons this (Plus generally, and this solution specifically) didn't work. But I wish it or something like it had.

Google's failure vis-a-vis Facebook convinced everyone Facebook's theory of the Internet (or at least social media) was right and Google's wrong. But I think this article, and Facebook's performance the last several years, shows otherwise, to the detriment of all of us.

Expand full comment
Dec 16, 2022Liked by Noah Smith

Worth pointing out the effect that Twitter had on people of middling renown. It's obvious that the super famous and connected would grow rapidly on Twitter - but it also rewarded people who had the time to post a metric ton on Twitter. Now - I'm not saying that someone can't be an expert or carry a high amount of domain specific knowledge and post on Twitter - but they are inversely correlated. It's sensible that the less specific knowledge you have - you are more likely to hit it big on Twitter in terms of notoriety (the effect I'm discussing is marginal at best to be clear.)

The question is - do we want to listen to people like that? If so, why? It makes sense that people that know a lot about cars would share knowledge about cars on a cars forum - but they would never take over Twitter. They literally can't - it's not generic enough information. Should we be consuming generic-grade information, or worse yet, elevating the ideas of people who have lower quality ideas than you would expect given their stature on the platform? The most intelligent and knowledgeable people I've met in my life are business owners, skilled and intelligent workers, and college professors. I've learned a lot and had mind-blowing conversations from mostly just listening to them and asking questions. Exactly zero of these people are on Twitter.

Expand full comment
Dec 16, 2022Liked by Noah Smith

Love this. You should include the need for ambassadors to help different towns communicate peacefully and profitably.

Expand full comment

I would add deracination to fragmentation. The ability to plant and nurture deep roots seems less possible or at least less apparent on TikTok, Instagram, and Snapchat. I've been thinking about T.S. Eliot's 1919 "Tradition and the Individual Talent" as a once relevant theory on how the 'best' creators deliberately situate themselves in relation to the best of what has been created. Without an archive how is that possible.

Expand full comment
Dec 17, 2022Liked by Noah Smith

On Twitter, different communities tend to stay separate. Law Twitter, or medieval Twitter, or whatever are often having their own separate discussions. It's mostly on contested, national political issues that become people shouting at each other.

You mention allowing people to moderate responses to their tweets. I argued that Twitter should allow that. I think it would enhance the ability of people to curate conversations and communities in their corner of Twitter. Some would choose to delete anyone who disagreed with them. Their comment threads would be boring. Others would choose to allow dissent while limiting the shouting.

https://twitter.com/jmarkgilbert/status/1594432471902617600

Expand full comment

Agree! I have high hopes for Mastodon because it is based on labor of love moderation, communities that set rules for themselves, and wide choice of communities. That said, its federated model still needs work. Cross-community communication has to improve and it has scaling issues that cry for attention from engineers who understand cloud scale and network engineering. I wish something like the Linux Foundation or Apache would form to take over engineering the platform and poach a bunch of AWS and Azure engineers to make it all sing. Also, grow a support infrastructure for community moderators, like the ITIL training and support industry with a lot less bureaucracy. I am forever hopeful.

Expand full comment
Dec 16, 2022·edited Dec 16, 2022Liked by Noah Smith

Like you mentioned, it seems like currently thriving platforms have powerful content recommendation. It feels like the move towards self-segregation is almost a byproduct of this, rather than an intentional choice by individual.

Or maybe the popularity of platforms is just part of the Boom-Bust cycle. /shrug

Expand full comment
Dec 16, 2022Liked by Noah Smith

"That experiment failed. Humanity does not want to be a global hive mind. We are not rational Bayesian updaters who will eventually reach agreement;"

Not even Bayesian updaters need reach agreement when given the same new data, since the same new data will update priors differently based on what the priors already are.

From p 157 of this PDF:

http://www.med.mcgill.ca/epidemiology/hanley/bios601/GaussianModel/JaynesProbabilityTheory.pdf

"From this we might conjecture that, whatever the new information D, it should tend to bring different people into closer agreement with each other, in the sense that

|P(S|D I A) − P(S|D I B )| < |P(S|I A) − P(S|I B )|. (5.20)

Although this can be verified in special cases, it is not true in general. Is there some other measure of ‘closeness of agreement’ such as log[P(S|D I A)/P(S|D I B )], for which this converging of opinions can be proved as a general theorem? Not even this is possible; the failure of probability theory to give this expected result tells us that convergence of views is not a general phenomenon."

Expand full comment

This is spot on. I feel like we are going through the tumultuous adolescence of social media and will come out of it the wiser over the coming years and decades.

This may be an effect of the pandemic, but the average amount of time spent on devices started to drop this year for the first time ever. Not only are people moving towards more niche outlets, they are moving more offline (getting the proverbial break from internet): https://datareportal.com/

Hoping to write more about this soon!

Expand full comment
Dec 16, 2022Liked by Noah Smith

How would you factor in Tumblr in all of this? Not being a participant there, my sense was that things went bad there before they ever did on Twitter and FB, and in a similar but more exaggerated way.

Expand full comment

I like this idea in theory, but how do you address the issue of network externalities?

Most social media companies don't actually bring a whole lot of "tech" to the table. They're successful because they gained first-mover advantage in a particular space where network externalities are strong.

Twitter is an extreme case. It does use human employees for content moderation and algorithms for prioritizing what you see in your feed (although I've turned off the latter, as everyone should). But if you set those aside the company doesn't really "do" much of anything. Its value to each user consists mostly of being the place where other users are.

I think this is why the debate over "free speech" on social media is so intractable.

The ideal libertarian model is one where the actions of private entities never have to count as infringements of free speech, because they're entitled to make their own rules. If one publisher won't touch your controversial book, you go to the publisher next door.

The problem with this model is that it assumes perfect competition in publishing. When there's significant market power, decisions by private actors are quite capable of making people feel censored.

This is one reason we used to have the Fairness Doctrine and other rules for the broadcasting sector. In the days when there were only three television networks, it didn't seem like "free speech" to give those companies total control over content, so some regulation was accepted.

Those rules are gone now and there's not much reason to bring them back, since the internet has eliminated natural monopolies in content delivery. (Except at the ISP level, which is why people want net neutrality enforced by law.)

The problem with social media, though, is that network externalities create a permanent likelihood of extreme market dominance even where there's no natural monopoly (in the sense that economists use the term).

A "permanent likelihood of dominance" isn't the same as a likelihood of permanent dominance, because firms can stumble and lose their position. (Just look at Facebook.) But I don't think that improves the situation very much. The long-term outlook is still that a few huge companies will dominate the public discourse and that their content moderation rules will shape the discourse in unaccountable ways.

So I don't really see how Noah's fragmentation model can solve the problem with Twitter.

Remember: Twitter is *already* a niche product, with many fewer users than the market leaders. How much more fragmentation could a Twitter-like entity survive?

The reason people have become so upset about its internal rules, first before Musk and now under Musk, is because although it's a niche product, it happens to be a niche product *for journalists*. (And perhaps more importantly, for laypeople with the mentality of journalists. A lot of us want a social media app that helps us to understand many things at a general level, not to gain deep expertise on one specific topic.)

I just don't see how multiple mini-Twitters, each with its own rules, could address the dissatisfaction people feel about Twitter itself. If each of those spaces covered a narrow range of subjects and *also* made its own moderation rules, people with an interest in those subjects would still be disgruntled if they didn't like the rules.

And everyone with a broad range of interests--the journalists and quasi-journalists--would be disgruntled because there would be no single place to go for every subject, just a bunch of random Reddits or RSS feeds. (Not to mention that if we had the patience to follow all of those feeds individually, we'd immediately get annoyed with each feed whose moderation policies we didn't like.)

I don't have a solution to propose. I'm just suggesting that this is an unsolved problem of market failure.

Expand full comment

The internet went to hell in 1993, the “year of endless September” when the old Usenet standards which had been passed down and enforced by the older users, got overwhelmed by the rush of newcomers from AOL. Though certain well curated sections, like this comment section, approximate it sometimes.

Expand full comment

You are a genius, sir. Always just stunned at how cogent and mind-boggling your analysis is.

Expand full comment

Some of the points you make about “filter bubbles” seem ripe for Thi Nguyen’s distinction between true “filter bubbles” (where you don’t see the people who disagree with you at all) and “echo chambers” (where you see the people who disagree with you and learn to hate them and distrust them). His point is that echo chambers are more pernicious than filter bubbles, and this lines up with what you say about the fragmented internet.

https://aeon.co/essays/why-its-as-hard-to-escape-an-echo-chamber-as-it-is-to-flee-a-cult

Expand full comment
Dec 16, 2022Liked by Noah Smith

Much insight of what is obvious when it is written down. It’s true - we need separation.

Expand full comment