Friday, December 09, 2016

Is Twitter a dystopian technology?

"What will the apocalypse look like? The answer, to use a term generally understood but the specifics of which you cannot imagine, and which this document will attempt to describe, is 'warfare'."
- William Bell, Fringe

I've been wondering whether Twitter is a true dystopian technology. Meaning, a technology that makes each user better off, but makes the world worse off as a whole. 

Note that you don't need a dystopian technology to create a dystopia. Joseph Stalin, Mao Zedong, and plenty of others managed to create dystopias using much the same technologies used in much happier, freer places. The question is whether some technologies, merely by existing, push the world toward a bad equilibrium.

How could a technology that's good for each individual be bad for the world? Externalities. For example, suppose that the only fuel we had was so horribly polluting that it would destroy the planet if we used it for even just a few decades. But each individual's choice of whether to use the fuel wouldn't be enough to tip the balance away from or toward planetary destruction. So without some kind of outside authority banning the use of the fuel, or some massive unprecedented outpouring of altruism, the world would be doomed. In fact, if global warming does destroy human civilization, fossil fuels will turn out to have been a dystopian technology.

What are some other examples of dystopian technologies? I'd say the jury is still out on the nuclear bomb. If nuclear deterrence dramatically reduces war and no one ends up using nukes, then it was a good technology. But if nuclear war eventually blows up civilization, it was a dystopian technology. Obviously, if weapons of mass destruction got cheap enough, they'd put society in untenable danger. I'm pretty sure that the ending of The Stars My Destination, in which (SPOILER) a guy runs around tossing out weapons of mass destruction to everyone on the street, won't turn out well.

Another possible example might be the stirrup. Stirrups enabled cavalry to become peerless warriors. For at least a thousand years, cavalry ruled the battlefield, and nomadic tribes from the Huns to the Mongols raided and/or conquered every settled civilization. Not coincidentally, during this millennium (usually called the Middle Ages), human wealth and development didn't advance very much. Then again, when global trade eventually linked the world, horses were an important piece of the equation. So without knowing whether globalization and science would have been possible without stirrups, it's hard to tell whether stirrups made the world worse off. 

These examples illustrate how hard it is to do a cost-benefit accounting for a technology, even with the full benefit of millennia of hindsight. So I don't expect the question I ask in this post to ever have a satisfactory answer. But in any case, it's interesting to ponder.

Is Twitter a dystopian technology? Obviously, Twitter is worth it for hundreds of millions of individuals to use. The costs - harassment, bad feelings, the risk of accidentally tweeting something that harms your career - are considerable, and no doubt have contributed to the company's user growth stagnation. But for many users, those costs are still worth paying. 

But what about the externalities of Twitter? Well, one of the biggest negative externalities is war. Twitter's format is very conducive to firing off quick thoughts without carefully considering the consequences. The brevity of tweets also makes them easily subject to misinterpretation. Imagine if President Trump fired off an intemperate tweet that leads to a nuclear war with China or Russia. It's quite possible that if he were forced to use a longer-format medium like Facebook, the risk of that happening would be much lower. Already Trump has used Twitter to denounce China's military buildup in the South China Sea.

So there's that.

A more insidious way that Twitter could generate negative externalities is to contribute to political polarization. In an earlier post, I described why Twitter is conducive to bad feelings, aggression, perceived aggression, and the creation of aggressive online mobs. If you don't believe me when I say Twitter is a vicious jungle, look at what happened to the Microsoft chatbot that taught itself to tweet by watching other people:

In brief, the features of Twitter that make it conducive to fighting are:

1. Limited length of replies. 

This is the big one. Like the faster-than-light internet in A Fire Upon the Deep, Twitter allows only short, pithy replies. Short replies have no room for self-deprecation, qualifiers, jokes, praise, caveats, or any of the other social lubricants that make debates friendly and collegial. Typically, the only way to get the point across in 140 characters is to be blunt. 

In addition, short replies make it harder to establish a voice when writing; that forces readers to project an imagined voice onto each tweet. Some people will assume the best, but others will assume the worst - that the writer of the tweet is being rude, sarcastic, or aggressive. And since an (assumed) unfriendly tone does more social damage than an (assumed) friendly tone heals, the net result of random errors in tone-interpretation will be to create bad feelings, including resentment, offense, threat perception, and a feeling of persecution.

2. Retweets that make replies impossible.

When I retweet someone, all of my followers can see both her tweet and my added commentary. But the author of the tweet cannot issue a reply that all of my followers automatically see. She can reply, but only if the readers scroll down through the thread will they be able to read her reply. 

So imagine if someone tweets "I think Democrats should try to appeal to some Trump voters," and I quote this tweet, adding the comment "This person thinks we should coddle racists." All 46,700 of my followers can see my uncharitable interpretation of her statement. But if she wants to reply "No that's not what I meant at all," then my 46,700 followers will only see her reply if they click on my tweet and scroll through the thread. Most will probably not do this, since it takes time and effort. Instead, they will probably see my uncharitable interpretation, assume it's true, and feel contempt for the person who wrote the original tweet. Also, my followers will now know her handle, so they will be able to tweet mean things directly to her ("Asshole, please delete your account", etc.).

I see this happen all the time.

3. Open mentions.

On Facebook, only people I approve can comment on my posts. I can make posts open, or I can limit replies to only my personal friends. On Twitter, on the other hand, anyone can automatically reply directly to anyone else, unless they have been blocked or muted. Since mentions are the main way that people talk to each other on Twitter, this means that if you want to talk to people on Twitter, you have no choice but to scroll through the replies of anyone who decides to talk to you. There's just no way not to. In addition, there's no way to untag yourself from other people's tweets, as you can do on Facebook.

4. Anonymity

Twitter, like discussion forums - but unlike Facebook - thrives on anonymity. This of course makes the platform have lots of bots. But more importantly, as we all know, online anonymity allows people to blow off steam, and reduces people's incentive to be friendly and diplomatic.

These three basic features of Twitter's technology encourage aggressive discourse, bad feelings, harassment, and constant rhetorical combat. That's certainly a cost to the user, but it might also encourage partisanship. Each person, in order to feel emotionally safe from the constant attacks that he feels like he's getting on Twitter, might be pushed to join an ideological group - like a prison gang, for protection (this excellent analogy was originally made by Charles Johnson, the nutty right-wing reporter).

Ideological polarization creates few costs for the user. It really doesn't make my online experience much worse to join the BernieBros, or the Alt-Right, or the Social Justice Warriors, or GamerGate, or the Libertarians, or whoever. I sacrifice a little bit of opportunity to say maverick, unorthodox things, and in return I get a whole bunch of people who have my back and are willing to beat off waves of attackers on a daily basis. 

But ideological polarization might be very costly for society. Eventually, if no one listens to the other side, people can come to believe that everyone in the opposing tribe hates them and wants to destroy them. The result can be large-scale social strife, civil war, or simply long-term political dysfunction. 

Many people have remarked that Twitter wielded a huge amount of influence in the 2016 presidential election. That election generated unprecedented level of bad feelings. You can chalk this up to the candidates themselves, but it seems very likely to me that the constant, unending, bitter Twitter wars were a big part of what made the campaign so unbearable. I myself would occasionally "detox" from campaign Twitter for days at a time, and find myself feeling much better about U.S. politics, about both candidates, and about life and the world in general.

But note that political tribalism can also dramatically raise the cost of quitting Twitter. Even if the online conflict is grueling and unpleasant - and even if it would make everyone happier if the conflict would just stop, or at least quiet down - one can't abandon the battlefield to the enemy. Once you join and commit to a Twitter tribe, you can't just check out, any more than you can desert your platoon in the middle of a war. 

So it's possible that Twitter's existence is adding significantly to America's already worrying polarization problem - causing American society to turn into a new sort of war zone. And if this is indeed the case, it's the 4 aforementioned basic features of the technology that are doing it. 

Maybe the people who run Twitter are smarter than me, but personally I just don't see a way to get rid of any of those 4 features without fundamentally changing the product. And this is a product we know makes money, has a strong network effect, and fills an important niche in the media. Twitter's problems seem inherent to any Twitter-like medium, and society seems like it will always have a need for a Twitter-like medium as long as it's technologically feasible.

So to sum up: Twitter is a great tool, and creates lots of social value. Personally, I have no intention of quitting, and most of the existing users don't seem likely to quit either. But it's possible that by exacerbating polarization and fostering large-scale social conflict, Twitter will end up destroying more social value than it creates.

If this is the case - if I haven't exaggerated Twitter's social costs or underrated its benefits - what's the solution? The government could just ban Twitter, or heavily police it like in China. That seems like it would have very high costs for society, since a government that does that has basically chucked free speech out the window (just look at China). So I think that society's only real way out of Twitter Hell is to innovate its way out. Solve a dystopian tech problem by inventing utopian tech - a new product that fulfills the useful social role of Twitter but avoids the negative externalities, by somehow fostering peaceful, friendly, constructive dialogue. Or modify Twitter in ways that fix its problems, which of course I haven't thought of yet. If stirrup-using horse nomads keep conquering civilization, invent the gun. 

Which is basically a cop-out. "Invent magical new technology that solves the problem." Check. But since the march of technology only goes forward, if you stumble into a dystopian cul-de-sac, there's generally no way out but forward into the unknown.


  1. The easiest way to fix Twitter would be for it to take on some Facebook-style attributes, with lots of tools for curating follower counts and your interaction. Just being able to mute hashtags, for example, has been great - and Twitter does let you effectively create "private" communities on the site with protected accounts.

    Not coincidentally, during this millennium (usually called the Middle Ages), human wealth and development didn't advance very much.

    Minor nitpick, but that's a myth.

  2. The Greater Internet Fuckwad Theory is a critical insight into our society that I invoke on a regular basis!

  3. Agree about twitter needing to evolve.

    Or I as put same thing (somewhat sardonically) a few days ago on twitter:

    Fermi Paradox Stages
    1. unicellular life
    2. complex life
    3. social animals
    4. tools
    5. invent twitter and fake news: self-destruct

  4. $5 tax per tweet. Solves the polarization problem and the deficit simultaneously.

    1. Best idea I have heard in a long time. Like a $10 figure better.

  5. 2 and 3 seem easy to fix. Quote retweets could automatically be included as replies to the original thread, and more nuanced privacy options would not be difficult to implement.

    1 is the magic of the medium, I wouldn't mess with it, though I do think Trump has caused enough trouble to deserve his account locked!

    And I have no good ideas for fighting polarization yet, but better privacy controls could limit the scope of the flame wars.

  6. "And this is a product we know makes money..."

    No actually we know it doesn't.

  7. I think Twitter is exposing the existing rift but not causing it.

  8. "Twitter is a great tool, and creates lots of social value" I think you should provide some evidence for that statement.

    Counter-evidence: Krugman used to write about three times as many interesting little blog posts a week. His tweets don't have nearly the same value.

    1. PK is a perfect example of what Our Host is saying. He had a million followers on Twitter before he'd ever tweeted a word. If you're a public intellectual, how do you pass up an audience that big?

  9. Twitter is perfect for the Sesame Street raised individuals. A subject change lacking substance every 30 seconds.

    1. I resent this! I was raised on Sesame Street (in the 1970's, so the original cast), and I'm all about long term focus, I read actual books, and find Twitter aggravating and panic-attack-inducing. I urge you, sir, to withdraw your calumny against Sesame Street.

  10. The rationale for the U.S. and Russia to have thousands of ICBM nukes on hair trigger alert isn't to deter war in general, it's to deter nuclear war. So no, nukes are not making us safer.

    (Not that "civilization" is at any serious risk. It's based on written information, which can't be destroyed as easily as people and buildings).

  11. Why I don't have a twitter account, why I don't follow anyone on twitter, why I avoid twitter as much as possible...

  12. "Each person, in order to feel emotionally safe from the constant attacks that he feels like he's getting on Twitter, might be pushed to join an ideological group [...]"

    Why not just ignore it? If people are saying nasty things about you on Twitter, walk away and spend the time on something more useful. It's not as if Twitter were food or oxygen, something you can't survive without.

  13. What you describe are the downsides of a networked society: in a village society you have to get along (more or less) with everyone because you may need them tomorrow; in a network everyone is free to engage only with people who agree with them, thus creating endless self-reinforcing loops leading to greater and greater polarization. So we have a tool that reinforces these tendencies AS WELL AS supporting, to some extent, the positive sides of networking. Sure it could be improved - Good points, Sarah.

  14. Noah, you used way more than 140 characters in your Twitter explanation! I've never signed up for Twitter as it appears to be yet another app that can suck the time out of the day (much the same way that Panda Pop does) and one gets preoccupied with Tweets. I read Tweets passively when they are part of a mainstream news story or a blog such as yours. I don't feel that I am missing anything.

    I would add one final note, given the President-Elect has cancelled (rescheduled???) yet another news conference yet continues to Tweet. It's likely that the only news about President Trump will come from Twitter as it's increasingly unlikely that he will respond to the mainstream press.

  15. In many ways, Twitter is the CB radio of this generation. Whoever has the most powerful platform can use it to lecture, insult, and shout down opposition, (regardless of whether either party is well-informed.)

    The next-level abuse of Twitter will be the "Henry II" variety, where the powerful tweet out the names of people who are troublesome, and followers set out to rid the world of them in the manner of Thomas Becket.

    Trump followers have not yet taken it to the "St. Thomas of Canterbury" level, but there have been many colorful threats to do exactly that. Can we say right now that no one will ever act out those threats?

  16. Just as unregulated "free" markets destroy themselves in monopolies, unregulated social media destroy themselves in chaos. In pre-internet times, paper media were self-limiting by the cost of owning a printing press, the tedium of typesetting, and the costs of paper, ink, and distribution, and electronic media by the cost of TV/radio production facilities and limited broadcast coverage areas, and those limits created a fairly high quality/cost threshold. Those limits are all gone now, so artificial limits are required in order to obtain continued survival.

    Facebook et al. have begun to institute things like enhanced moderation and fake news suppression, but due to its founders' misguided ideals, Twitter is probably going to be among the last to raise its game to the needed level. It may well end up like Yahoo comments or 4chan, a wretched hive of scum and villainy that is so popular that nobody goes there any more -- "nobody" being paying advertisers.

    China has had an obsession with civil order since Confucian times, and has the vast manpower to impose that order on its internal internet -- Sina Weibo is in no danger from Twitter's troubles. I don't know if America has the creativity or discipline needed to keep the instability within the bounds needed to avoid catastrophe. Our Bully Pulpiteer certainly doesn't have that discipline, but the forces against net neutrality may actually create barriers to entry that keep the most volatile forums out of the mainstream, Breitbart notwithstanding.

    On the other hand, this conventional wisdom could be totally wrong, and social media actually reduce polarization. But my mind is made up, and I don't want to be confused by empirical facts like the ones in this paper:

  17. Anonymous3:29 AM

    Why can't we regulate internet communication but retain the free speech we had in 1990?

    Freedom of speech would still exist if the internet disappeared tomorrow.

    We regulate access to TV broadcasting licenses, to radio. Why not regulate the right to post information on the web where anyone can see it with a Google search?

    Why should anti-vaxxers be able to multiply the number of kids they make sick using technology which can only operate with social consent?

    I don't think a hand wave about free speech is really a good enough answer. We are talking not about free speech, but about broadcasting rights.