top of page

Sideways Looks #25: Musk, Micro, Macro

Updated: Sep 12, 2023


Back in the distant past (May) I wrote in Tech Fought The Law, and The Law Lost:

As you may have heard, there’s been a bit of a saga about Elon Musk buying Twitter … given a few people have asked me for thoughts: I think a lot of the most concerned takes are overblown; I do not think Musk owning Twitter would turn it into a massive far-right hate speech space, akin to Gab or Parler but much bigger. Some rough stats: According to Twitter, they took down a few million tweets and accounts in the 6 months Jan-June 2021. By comparison, in 2020 there were apparently 200 million active users and 500 million Tweets per day. So ... there’s a long way to go before hate speech makes up a substantial chunk of Twitter.

What Musk owning Twitter could quite likely do is make the experience considerably worse for a marginalised minority of users, who will now have even less confidence that they can get redress against abuse (and might face even more abuse, if Twitter becomes seen as a space which welcomes unpleasant forms of “free speech”). It’s worth remembering that things on Twitter aren’t always great for such users right now. But at least it was, maybe, trending in the right direction.

Separately, Musk might push for unpredictable tech changes to the platform, many of which would probably be directed at making money and might make the experience of Twitter quite different and maybe worse (e.g. paying for verification).

Also Musk is a man with a deeply worrying approach to treatment of humans, and I’d rather such people just owned less stuff in general.

A few weeks into Musk’s tenure as Chief Twit (his own label) I still stand by these views. But to clarify - I’m not saying we shouldn’t be concerned. Rather I’m saying that a lot of the concerns might be focussing at the wrong level.

What do I mean by ‘the wrong level’? Well the key ideas above, as you can tell because I’ve put them in bold, are “massive” and “minority”. Social scientists might call this “macro” and “micro” levels.

A lot of commentary about Twitter has suggested large macro-scale impacts - Twitter will die completely, the whole site will become a hellscape, it’ll help the far-right gain more influence in politics, etc. The risk: if that doesn’t happen, lots of people might say ‘wow this doesn’t seem as bad as expected’. From that they'll conclude everything is basically fine - that Musk’s critics were exaggerating, that ‘the liberals’ are panicking again, etc.

But if you look at the micro level, you’ll see problems emerging for specific people, groups, and events - they just won’t be writ large on everyone’s Twitter experience. And micro effects can still cause big problems.

So let me draw that out by taking you down a funnel from the macro level, steadily down to the micro. So to start very macro…

How might Musk’s acquisition of Twitter change the world?

That’s a big question. It’s actually a subset of an even bigger question - how does social media change the world? Which is very complicated to address. I know because I tried to start with that question; which spiralled out of control, so it’ll be a post of its own.

The short version relevant to today is: there’s a lot of claims that social media has changed how society, at large, thinks and behaves. Social media broadcasts content to a mass audience, which has created a polarised, echo-chamber-ised, fake-news-riven world. But these claims are not clearly supported by evidence; this book is a very sober, but very useful, summary. In fact, social media arguably gives people more visibility of alternative viewpoints (so helping us see beyond echo-chambers), and also makes fact-checking faster, than a world where information is largely controlled by the mainstream media (particularly the Murdoch media empire). Also, it seems likely that the complex, unequal, and tense modern world was always going to be polarised.

I’m not arguing social media is unimportant. I’m just saying we should treat claims of social media’s revolutionary influence on society’s general mindset with a bit of scepticism.

Whatever their influence or not, it is true that the content of Facebook, YouTube, and increasingly TikTok can reach 100s of millions of people - even billions - of people. But what about Twitter? Well, have a look at this graph.

(That graph is for the USA; but the general patterns are reflected in UK data, from Ofcom).

Look at how low Twitter is. Hanging out below Pinterest and Snapchat. When was the last time you saw a journalist or politician talk about either of those?

The columnist Danny Finkelstein once remarked that “history isn’t written by the victors, it’s written by historians”. His point: the way the world gets written about tends to be biased towards the interests of, well, the kinds of people who get paid to write about the world. An important part of Twitter is that it has absolutely captured the attention of those kinds of people. As the tech journalist Nilay Patel put it in his excellent piece 'Welcome to hell, Elon':

The asset is the user base: hopelessly addicted politicians, reporters, celebrities, and other people who should know better but keep posting anyway. You! You, Elon Musk, are addicted to Twitter. You’re the asset. You just bought yourself for $44 billion dollars.

But Twitter’s influence comes precisely from the fact that it is the place where specific people, groups, and events can - often unexpectedly - magnify their influence. Again, some of these stories are exaggerated. I said back in May that I thought the impact of Twitter on the political career of Donald Trump has been over-estimated; since then, I’ve encountered a study of the 2016 US election which suggests that Twitter maybe even hurt Trump’s vote a tiny bit, though I have. (So maybe it’d be good to let Trump back on!) But the use of Twitter by other specific people, groups, and events - including many outside the West - are perhaps more important. We’ll come to those later.

In the meantime, the answer to my question ‘will Musk owning Twitter change the world?’ is - well, it might. But if so, it would most likely be through its impacts on specific, unpredictable, hard-to-spot things, not big obvious trends. Again, more on that shortly.

For now, let’s step down a bit from the big macro-level scale - but not quite to the micro-level yet. If we’re being fancy we might call this the ‘meso’ level.

How might Musk’s acquisition change Twitter in general?

Hard to tell. Musk is a little, erm, experimental in his approach; also trying to turn a tanker of 450 million users will always be a challenge. But some possibilities are emerging from the maelstrom.

For instance, Musk’s suggestion that paying users would get content boosted over non-paying users - potentially such that you’d have to scroll a long time to see non-paying users - could make it a much less anarchic, random, folksy platform. It may also be less appealing to any creator who isn’t already influential. Or maybe all Musk’s layoffs mean the platform will get so spammy and glitchy that it won’t be fun to use. So Twitter just becomes a bit rubbish.

In fact, maybe Twitter will… just die? Comparisons have been made along the lines of “but no-one thought MySpace would die and it did.” But for Twitter to “die” would be different to MySpace. When people say that MySpace (or Friendster, or Facebook) have “died”, it’s because they do not have enough active users to function as a social network - i.e. people can’t just log in and know they’ll see lots of fun stuff from their friends. (MySpace tried pivoting to provide lots of other alternatives to social networking, but it didn’t work).

Twitter’s different. Most tweets are produced by a small number of people for everyone else to read (in fact, amongst adult U.S. users, the most prolific 10% create 80% of tweets). So as long as enough of its power users - the celebrities, the politicians, the journalists, and the actually good comedy accounts - stay, I suspect it’ll be basically the same experience. And I don’t see the power users leaving en masse anytime soon - and if they actually do, I suspect that would lead to some serious pivoting in Twitter HQ.

(Lots of people are exploring other platforms. That’s great. But there’s also been a bit of a meme that people who’ve been vocally tweeting about leaving Twitter… are still on Twitter. But that’s good, in my view. If people are concerned about Twitter, they should stay and help keep it healthy).

What about the other concern, that Musk’s approach to free speech could lead to Twitter becoming a far-right “hellscape” akin to 4chan, Parler, Gab, or suchlike? I've been involved in research with CASM Technology and the Institute for Strategic Dialogue which used various methods, including AI-based Natural Language Processing and analysis of newly-created accounts, to demonstrate increases in antisemitism and misogyny, and some other research even showed the appearance of more Islamic State accounts. Though there's also been some less good research, some of which simply counts increases in uses of particular words (which doesn't account for context), and others including some BBC research which did the rounds recently using an AI tool in which “you are a poo poo head” was classified as more toxic than white supremacy.

Such increases are is deeply concerning. But my original point from last May still stands. I think predictions that Twitter will become a “hellscape” are overblown; I don’t think enough of the people who actually post will change behaviour, or leave, to cede it to just hate-speech and conspiracy theories. This is particularly so given Musk himself seems to be cooling on full-throttle free-speech as many brands are now, unsurprisingly, expressing concerns that their ads will appear alongside hate speech - though who can tell what Musk really thinks.

(Separately - this latter point plus the downfall of Liz Truss are both being hailed by many, correctly in my view, as good outcomes. But both of those are cases where markets effectively decided political outcomes - and I think those of us on the political left need to think carefully before leaping too joyfully at such a context, even if we think they’re the right outcome).

But we can’t just stop there. Just because hateful speech doesn’t dominate Twitter, doesn’t mean it’s ok if it increases. That’s not just a moral issue - it’s because the micro level is still important.

How might Musk’s acquisition change Twitter in specific contexts?

Finally, this is the micro-level analysis; i.e. looking at the specific details of clearly-bounded cases. Where a macro-economist might look at large aggregated markets, a micro-economist might examine specific sectors and firms, sometimes in minute detail. The latter is a very important practice in a world where particular trade routes or commodities can have enormous impacts (as illustrated vividly by the Odd Lots podcast's recent obsession with trucking).

I argued in May that “What Musk owning Twitter could quite likely do is make the experience considerably worse for a marginalised minority of users, who will now have even less confidence that they can get redress against abuse.” Originally I thought that would be because of Musk’s ideological approach to free speech. I did not also anticipate the scale of layoffs and quitting, which further diminishes practical capability to deal with reports. This will be unevenly distributed across geographies and languages. In India - an enormous, culturally complex, and very politically fraught market - Musk has apparently fired over 90% of the staff.

That’s bad in its own right. But it also intersects with the point I mentioned earlier - using Twitter is increasingly becoming an expectation of elite figures. So there may well be potential politicians, journalists, and other influential figures who simply fail to materialise because Twitter is hostile to them. And they’ll likely come from marginalised groups, further impoverishing the diversity of our political and journalistic landscape. Maybe not; maybe they’ll find other platforms (Liz Truss was more of an Instagrammer than Tweeter). But getting into power tends to involve being in existing spaces of the already-powerful; and right now, that’s Twitter.

(A caveat: I think this is particularly true for journalists. One can become a prominent politician with little Twitter activity. But it still risks becoming another potential friction to new entrants).

Another micro-level issue to consider is specific events. Twitter helps protests, even revolutions, reach wider audiences. This is particularly key when authorities attempt to repress the flow of information - whether the Arab Spring in the past, or Iran and China today. Twitter is also a fantastically open platform with data, which makes it extremely valuable if you’re an investigative journalist, or open-source researchers trying to quickly establish reliable facts about breaking events (e.g. the war in Ukraine). This openness has perhaps meant too much ‘general’ research is done on Twitter; but for fast-moving events it really is a fantastic resource. All that effectiveness could diminish if Twitter gets glitchier, or spammier, or non-paid-for content becomes harder to find, or if deception becomes easier - or even just because of Musk-induced unpredictability amidst some geopolitical event.

(I recently did research with The Institute for Strategic Dialogue on how extremists splintering away from mainstream platforms can make researching - and tackling - them harder. This may now become an issue for researching social media in general).

None of the above would appear as a big bang macro- or meso- level collapse of Twitter: the platform dying, or a mass exodus of power users, writ large across everyone’s Twitter experience and making headlines and commentary pieces. It would be a boiling frog, a slow accumulation of specific degradations - political and journalistic careers which don’t happen, events which escape wider scrutiny. All that ultimately, might, still have an impact on the wider world. We just won’t necessarily see them happening.

So that’s why I’m concerned

There are lots of great people already highlighting the specific issues I’ve just outlined (follow Joanna Bryson and Eliot Higgins for starters). But back to my original point: the voices calling us to look at Twitter at a macro level - both concerned critics and boosterish supporters - have been loud and prominent, and risk unhelpfully setting the terms of debate.

A case in point. Musk recently claimed that “impressions” - i.e. the number of times a tweet appears in people’s feeds - of tweets using hate speech had fallen by ⅓ (using internal Twitter data, so...). I.e. even if there's more hate speech, it's not being seen by as many people. But not all "impressions" are equal. If a promising young, female, black journalist or politician is being bombarded with hate speech, with no confidence of support from the platform - is it still ok that the larger Twitter audience aren’t seeing as much of that hate speech? She's still seeing it. It's targeted at her. From accounts now welcomed on the platform, with little fear of reprisal.

(I’m aware I’m skipping over some big debates about content moderation. I’ve written more on my own views about making social media more pleasant and diverse, for the Tony Blair Institute).

In conclusion - I and Musk (apparently) share a hope that Twitter can be some sort of ‘public square’. My concern is not that the public square will collapse, or turn into a mass brawl. It’s that, slowly, people who brought diverse opinions and rigorous fact-checking will fade away and won’t be replaced. The public square will seem vibrant; but we won’t see the people who have stayed away, the events we didn’t properly discuss, and the better public square that we missed.


Fun Facts About: German Christmas Markets


I was told Germany loves Christmas markets. This turns out to be very true. Indeed, Berlin alone has over 68 of them. And yes, there's a spreadsheet telling you when they begin and end. Some even run on Christmas Day.

My favourite so far has been at RAW Gelände, a derelict former train repair station (yes Berlin is sometimes a parody of itself). This is a medieval themed Christmas market, surrounded by wooden ramparts, complete with axe throwing (I was bad) and fire jugglers (I just watched). I was also introduced to Grünkohl, a wintertime dish which is sausages in a kale stew. I think I'll be sticking to Glühwein in future.




For the political nerds, the Westminster Insider podcast - hosted by Ailbhe Rea and Jack Blanchard - are consistently producing excellent long-form, historically informed, thoroughly enjoyable programmes on subjects from political comebacks to the Whips. For a less enjoyable but also well-informed take on current politics, I’d also recommend Sam Freedman’s Substack.

I’ve taken some long train journeys (Berlin to Hertfordshire via Brussels, Orpington, and Lewes) so I’ve entertained myself with some long form journalism; they’re also good for just curling in an armchair while hiding from December, long train journeys are optional. The Financial Times has had good pieces on the women in - and leading - MI6, the last days of Liz Truss, and the bizarre recent far right plot to overthrow Germany; the New Statesman had a wonderfully written piece on Prince William, and a real page turner* on the car crash of the first year of GB News.

* Or whatever the audio equivalent is, given I listened to it as a podcast. Also the FT pieces are paywalled but I can send gift links.

Finally: The play Lemons Lemons Lemons Lemons Lemons is having its first run in the West End in January. It is superb. I originally saw it when it was a student production; I know the playwright, Sam Steiner, as we played the blood brothers in Blood Brothers in a youth drama club (I was the scouse brother). But it turned out the play, and Sam, have gone on to be mega-hits. He’s gone on to write for Benedict Cumberbatch. But I’ve gone on to write for all of you, so I reckon I’m winning.


Thanks for reading. Please do share with others, and let me know your thoughts via this short poll.


89 views0 comments

Recent Posts

See All
bottom of page