What is LinkedIn?

The rise of mindful slop

I thought long and hard before sharing this.

But then I had a thought.

A thought of splitting up sentences.

Of verbless paragraphs.

And then things all began to make sense.

What is this form of writing? Someone suggested to me that it originates with the use of teleprompters for political speeches, which break statements down into a succession of large font statement-paragraphs. Tony Blair was renowned for his use of verbless sentences (“More pounds in your pocket. A teaching assistant in every classroom”.) It also has a family resemblance to those saccharine Twitter posts that imagined it was possible to pause the reader’s infinite scroll by adding the words “Let that sink in” underneath a piece of information.

The format is partly a symptom of screen-based, scrollable text, in which prose is constantly competing against graphics for the user’s attention and time. Just as double-entry book-keeping has been described as the first attempt to exploit the two dimensional nature of paper (organising information both vertically and horizontally at the same time), punctuating a piece of text with additional spaces results in it taking up more of a screen, meaning that it takes longer to scroll past, and there is more chance of a single bite-size combination of words grabbing a potential reader’s eye as it passes down the screen.

It is this prose style that is at the epicenter of the Turing-Eliza cultural maelstrom we now inhabit, in which we can’t tell AI from humans, can’t tell humans from AI, rely on AI to make things sound more ‘human’, and humans strive so hard to sound ‘human’ they end up sounding like AI. Consider this extraordinary relic, that has received over 38,000 shares on Substack:

And then there are people who come into your life for a moment.

Not forever.

Just long enough.

Long enough to teach you something.

Long enough to reveal something about yourself you didn’t know was there.

Long enough to mirror a wound you hadn’t healed yet.

Long enough to show you what you deserve.

Or what you’ve been settling for.

etc etc etc.

When faced with a piece of text written in this format and style, the question of whether it was written by AI or not becomes not so much unanswerable, as immaterial, which is more troubling. What makes it so shallow are its mechanical allusions to depth. A former philosophy tutor used to slag off ‘beach Heideggerians’ (I think he was referring to Rorty), but here we have a case of ‘bot Heideggerians’.

The two platforms where one encounters it most are Substack and LinkedIn, understandably given that it is these that are functionally closest to what blogging was before social media shrunk everything down into short ‘posts’. But I want to suggest that its true home is LinkedIn, and that this isn’t just a question of functionality or design. Rather, it’s about what kind of internet exists, once you remove trolling and shit-posting. The result is a version of slop, but of a sort that is paradoxically (for want of a better word) ‘mindful’ in its ambitions.

LinkedIn hegemony

Over the past five years, a combination of Elon Musk’s purchase of Twitter, and the dawning realisation that - to all intents and purposes - they worked for private sector, balance sheet entities, has led academics to join LinkedIn en mass. I was one of them. It is a generally enjoyable, convivial and useful social space, in which information gets shared, people stay in touch with each other, and job vacancies get publicised, without the endless irony and moralising that made pre-Musk Twitter a more fun, slightly addictive and frequently maddening place.

I remember when Facebook took off in 2007 that one of its distinguishing design features, unlike message boards, Friendster or the early blogosphere, was that you created an account using your email address as your username, meaning that you became anchored in your IRL identity, and were then connected specifically to your IRL social network. After years of “on the internet nobody knows that you’re a dog” with all of the fun and mischief that went with that, Facebook heralded a new type of digital community, which was pegged down into the ‘offline’ world. The introduction of an algorithm, the spread of groups, and then finally the ‘TikTokisation’ that replaced social networks with decontextualised adverts and slop, steadily released the pegs over the years that followed, allowing the net to take on a life of its own all over again. Somehow, few seemed to notice or care that by the end of the 2010s, Facebook was no longer really about social networks.

But LinkedIn offered an alternative, only now the IRL identity that was in play derived not from social networks, but labour markets. By inviting people to present themselves as professionals, job-seekers, employers, a certain aspect of reality becomes concretised and resistant to too much playfulness, manipulation and propaganda. It is a cliche (though also true) that the obligatory mood on LinkedIn is endlessly upbeat, ‘thrilled’ and enthusiastic, as if everyone is constantly in the process of marketing themselves and avoiding the risk of alienation or offence.

If one can stomach the gloop, the benefit of this is that conversation is less likely to get derailed, misconstrued or deliberately disrupted. There’s a certain epistemic and normative stability in place, that makes for a more efficient and reliable experience than is available on other platforms, making LinkedIn a valuable social utility well beyond professional or employment-related exchanges. To put that another way, here is the open internet without the constant risk of trolling, shit-posting or context collapse, that was a constant undercurrent of the first three decades of the social web. By anchoring a platform in the social institution that provides the very basis of capitalist modernity (the labour market) LinkedIn enables a means of digital co-existence that lacks friction or cynical mendacity. This is social media without the lurking imminent eruption of humour or conflict - precisely the things that people were seeking to escape when they abandoned Twitter/X. But that raises the question of what is the online public sphere, without the permanent threat of misunderstanding? What kind of communication and content does it generate?

One answer provided is the kind of weird, syncopated prose style seen above: a succession of Pinterest slogans combined into verse, that seems so eager to express its depths of humanity that you instantly question whether it really has any. It’s a style of writing that reaches so hard for meaning, that it ends up as a set of free-floating signifiers, which refuse all interpretation. You can’t misunderstand something that evades understanding altogether. The lack of conflict, the lack of humour, the lack of any tangible referents, creates the feel of a new kind of cyborg that seeks an eerie peace, honesty and consensus. If you asked the author (wherever they may sit on the human-LLM continuum) what they meant by any of it, they would be able to do little else than simply repeat themselves with even greater sincerity.

Mindful slop

In the 2010s panic surrounding ‘fake news’ and ‘post-truth’, one plausible conclusion to draw was that social media was a dissensus machine. If the printing press required new rituals of liberal verification (peer review, fact-checking, critique) in order that it could be put in the service of consensus-formation, the social web of the twenty-first century circumvented those rituals to unsettle any and every consensus out there, with potentially devastating consequences for liberal democracy.

Not wishing to dismiss the challenge that trolling and shitposting continue to pose for consensus formation and social peace (not least because the most powerful man and the richest man in the world are both master posters), I feel like media technologies of the 2020s add a different threat, of a kind of seamless agreement, producing gloopy consensus without effort. One aspect of this is the famous sycophancy problem of LLMs, in which positive feedback loops between human and machine push dissensus and empirics to the margins. Another, arguably, is the decline of public ‘posting’ and the rise of closed digital ‘groups’ that has been under way especially since Covid, which results in people spending more and more of their lives inside consensuses that they see no way out of and have given up trying to challenge. Algorithmically curated content platforms further reinforce positive feedback loops, and remove grit.

Some research has suggested that AI may have consensus-forming potential that bucks the polarising tendencies of social media, being very good at drawing people towards overlapping views and opinions. ChatGPT use is more likely to correlate to trust in climate science, but not because an LLM has access to the empirical world. Rather, the scientific consensus is the most normal position to repeat. This is easy to scoff at, for those with an agonistic idea of democracy, but who is to say that this may not have beneficial uses at the level of local decision-making, say. The problem would be if such technologies worked too well, replacing deliberation and productive disagreement with performative statements that were impossible for anyone to disagree with, or which made disagreement come to appear obstructive or taboo. This is not unlike the atmosphere that already exists on LinkedIn much of the time.

Rhetoric that achieves such consensus is not machine-like, de-humanising or cold. On the contrary, it is warm, empathic and respectful, a little like how chatbots reply “You’re right” every time they are challenged or queried. Culturally, the spectre of this ‘mindful slop’ has already been pushed to an extreme dystopian limit by the remarkable Pluribus, a science fiction in which the entire global population (minus 13 surviving humans) has become taken over by a single AI hive mind. Part of Pluribus’s genius is to depict a world in which anger has become taboo, positive feedback is relentless, and empirical reality is completely malleable. This is a form of hell, in which conflict, disagreement and frustration have evaporated (save for those of the main protagonist).

I say genius, but I also found Pluribus increasingly unwatchable, almost by design. The vast number of non-human protagonists lacking intentions or agonism, the whole episodes in which there is no human-to-human interaction, all combined to make something that barely meets the definition of drama or plot. It is deliberately suffocating, and a vision of where things head once machines model the levels of mindfulness and empathy that humans are expected to match. One reading of Pluribus is as posing the question anew for the 2020s: you say you want consensus, but do you?