iiinova — where your digital self keeps evolving

Rob Horning @robhorning.bsky.social · 11/09/2026

"Rather than cultivating personal brands, they prioritise speed, scale and adaptability, pointing toward a post-creator mode of cultural labour in contemporary platform economies"

163

Rob Horning @robhorning.bsky.social · 09/09/2026

"All media consumption is reinforcement learning with human feedback" as a general description of the current cultural condition; no way of seeing now is exempt from it newleftreview.org/issues/ii160...

Advertising works as a social agent in this sense, as Paglen acknowledges, pointing to Edwards Bernays’s pioneering of public relations, and the early accounts of such activity in the work of Daniel Boorstin and Jacques Ellul. However, Paglen claims, ‘the efforts of the propagandists of yore are prehistoric by today’s standards, where the circulation of images is a largely automated circuit that reinforces itself through feedback mechanisms built into its systems’ so that ‘All media consumption is reinforcement learning with human feedback.’ The result is increasingly effective addiction, and precise tailoring to individual weaknesses as the boundaries between perception, reality and manipulation erode. Precedents for the extractive and controlling logics of machine imagery may be found at a small scale (psyops) and large (advertising) but its technology, reach, automation, processing of trillions of images and saturation of everyday life are entirely novel.

152

Rob Horning @robhorning.bsky.social · 02/09/2026

I guess the only reason you would call a data file an "image" is to make the claim that machines "see" rather than compute or process—but this essentially implies that "images no longer exist, everything is just data and all 'imagining' is just data processing" newleftreview.org/issues/ii160...

While the book is called How to See Like a Machine, it rapidly becomes evident that humans can do no such thing. Paglen tracks what he rightly sees as an epoch-making development, which is superficially marked by the recent inability of people to reliably distinguish machine images from those made by cameras or humans, and more profoundly the condition that has been reached in which ‘the overwhelming majority of images are now made by machines for other machines, with humans rarely in the loop.’ The computer file used to display a screen image may be thought of as a latent image, rather like an undeveloped photographic negative, unreadable to the human eye. Yet machines can read these files directly. Regardless of whether anyone attends to the two billion photographs uploaded each day to social media sites, machines scrutinize them for purposes of data mining, classification and manipulation.

272

Reposted by Rob Horning

👁️ @nathanjurgenson.com · 26/08/2026

a through-line from social media metrics to prediction markets: the fantasy of simple numbers that make us feel we know an unknowable world

271

Rob Horning @robhorning.bsky.social · 26/08/2026

prediction markets argue they are democratizing the risk-managing tools of insurance companies; this piece argues that is just a smokescreen for opening more tax-efficient sportsbooks thebaffler.com/salvos/infin...

While prediction market bets on the deaths of world leaders, the bombings of countries, or the burning of forests feel scandalous and obscene to many Americans, they are the direct descendants of those British underwriters betting on the sinking of ships or the starvation of helpless migrants—as well as the death bets of the viatical and life settlement industries. The plumbing of the American economy is corroded with all sorts of heinous shit like this, and all of it is at worst celebrated and promoted as prudent financial investments and at best completely ignored by the vast majority of Americans.

1226

Rob Horning @robhorning.bsky.social · 26/08/2026

prediction markets incentivize the production of insider information, which undermines the bureaucratic functioning of institutions harpers.org/archive/2026...

The prediction markets that exist today are not the fulfillment of Hanson’s vision; they are the fulfillment of May’s. Rather than coordinating expert opinions, they deliver uncertainty and distrust. Who’s cashing in on what they know? Who can be trusted? Is this story fake? Indeed, is any of this for real? Because of perverse incentives, a system intended to make information about the present and knowledge about the future more transparent has made the world more opaque and volatile. The dream of creating an honest consensus now works as a kind of acid, a solvent dismantling the capacity for coordination. What we’re left with is a system for producing conflict, a machine for generating chaos.

181

Rob Horning @robhorning.bsky.social · 21/08/2026

the gist of this post is that LLMs are terrible at drafting or editing text that anyone wants to read, but they are good at getting users to write text of their own — i.e. LLMs are serve an emotional rather than practical role in writing process www.patreon.com/MaxRead/post...

In general my feeling has been that the writing demanded by LLM chatbots is a benefit, not a burden.

1111

Rob Horning @robhorning.bsky.social · 12/08/2026

what Alasdair MacIntyre proposes as an outlandish thought experiment on page 1 of After Virtue describes what AI is doing to scientific practice right now

all that they possess are fragments: a
knowledge of experiments detached from any knowledge of the theoretical
context which gave them significance; pans of theories unrelated either to
the other bits and pieces of theory which they possess or to experiment;
instruments whose use has been forgotten; half-chapters from books, single
pages from articles, not always fully legible because torn and charred.

051

Reposted by Rob Horning

👁️ @nathanjurgenson.com · 12/08/2026

a good definition of "slop" might be the degree to which any output can only be judged quantitatively "Friction/scale is coming to replace low/high or pop/rock hierarchies" www.its-her-factory.com/2026/08/slop...

0115

Reposted by Rob Horning

👁️ @nathanjurgenson.com · 12/08/2026

AI as having a parasocial relationship to your own self and life

1258

Rob Horning @robhorning.bsky.social · 12/08/2026

LLMs present us with infinite "blocks of wood" masquerading as works mediated by thoughtful agency; these text-blocks impede our ability to "recognize ourselves" or others in language psyche.co/guides/how-t...

Imagine a sculptor faced with a block of wood. She works on the block with a set of awls and chisels, transforming it into an object of art, a statue. Through work, the original wood has been transformed. Its initial, heavy ‘woodness’ has been overcome. It has become an object of art invested with the sculptor’s thought, a reflection of herself, of her creativity. In Hegel-speak, the immediate block of wood has been mediated. It has been raised up and liberated from its initial form into something more mindful, more spiritual, more thoughtful. As such, it is something in which the artist may recognise or know herself as an artistic agent. More generally, our thinking agency transforms the world of objects, overcomes their immediate presentation and makes them into ‘objects of art’ in which we may recognise ourselves.

071

Rob Horning @robhorning.bsky.social · 12/08/2026

and of course current cultural logic points to ultimately having an "AI agent" do all the consumption for you while you get to check the boxes (as with the Axios worker who had AI process The Last of the Mohicans for him)

0122

Reposted by Rob Horning

slo ~|• @slo-slo-slo.bsky.social · 12/08/2026

“Those lines are drawn in terms of friction: devalued things like labor and politics require and generate friction, whereas scalable value must minimize it.” “In 2026, slop scales. Actual musical labor and the critical labor involved in interpreting music does not scale.”

065

Rob Horning @robhorning.bsky.social · 12/08/2026

now popularity metrics say nothing about social behavior, they measure the success of algorithms to atomize people with the same slop

052

Rob Horning @robhorning.bsky.social · 12/08/2026

"scale" as something individuals are (paradoxically) supposed to consume and enjoy is inculcated by all metricized platforms—the numbers reveal what is valuable. Numbers once were a proxy for social significance. Now scale means a stack of atomized consumption acts

000

Rob Horning @robhorning.bsky.social · 12/08/2026

this gets at the same idea with "sameness" as the lens: "Once prediction shapes the choices in front of us, we lose the ability to tell the difference between what people wanted and what the system made easy to want" — i.e. people consume "easiness" as the content, as the substance

070

Rob Horning @robhorning.bsky.social · 12/08/2026

commercial entertainment has been optimized for convenient consumption rather than "discourse," yielding "hits" that offer no substance for critical discussion etc. www.its-her-factory.com/2026/08/slop...

In 2026, slop scales. Actual musical labor and the critical labor involved in interpreting music does not scale. Critics dig their teeth into the products of musical labor, where the friction of work creates the open, messy, plurivalent art object ripe for the labor of interpretation (and the social proces of debate).

152

Rob Horning @robhorning.bsky.social · 07/08/2026

good account of how books and "book culture" offer no refuge from imperatives of streaming, parasociality, and "content creation" assets-us-01.kc-usercontent.com/f7ca9afb-82c...

assets-us-01.kc-usercontent.com

1104

Rob Horning @robhorning.bsky.social · 28/07/2026

"authenticity" is built from feelings not facts, and facts don't "feel real" in and of themselves. (also, "authenticity" is always a construct and not empirically given) ijoc.org/index.php/ij...

misinformation and disinformation do not proliferate merely because of the absence of literacy, trust, or discussions needed to discern facts from fiction. They succeed because a complex composition of actors ranging from social media platforms to armies of propaganda bots has
economized and politicized the affective texture of authenticity, hijacking the feeling of “real.”

071

Rob Horning @robhorning.bsky.social · 27/07/2026

LLMs turn our tendency to be considerate toward our interlocutors into a vulnerability rather than a strength; this makes for a more miserable world where we must be presumptively cruel and wary of anything that talks to us www.ft.com/content/bdb3...

Please use the sharing tools found via the share button at the top or side of articles. Copying articles to share with others is a breach of FT.com T&Cs and Copyright Policy. Email licensing@ft.com to buy additional rights. Subscribers may share up to 10 or 20 articles per month using the gift article service. More information can be found at https://help.ft.com/faq/gifting-and-sharing-an-article/what-is-a-gift-article/.
https://www.ft.com/content/bdb3b820-905b-431e-82c0-386535755af1?accessToken=zwAAAZ-kD1XYkdO9s7ggkFtDHtOCwDhlNXVa8Q.MEUCIQC-ETMCJ6SDtmAe3lji9VujA1qlZRfn_KjUZ70hHO5CAQIgJoYa2xerrzUVoelwDegvQirRES4ER26HFw6UudhU6RE&sharetype=gift&token=4976dabd-eafe-45cd-b702-1071a92607f4

After decades of marginalisation and perceived irrelevance, academics may not notice that they are being co-opted by a sophisticated PR machine. They are useful for adding intellectual credibility and pre-arming AI companies against criticism. With the resources at its disposal, the AI industry can buy anyone it wants, if not with actual currency then with commodities even more valuable to intellectuals: influence, prestige and recognition.

One form this co-option can take is keeping the focus of academic conversation on the moral and metaphysical status of LLMs. This is problematic in itself, but it also carries a devastating opportunity cost. It is much more important to conduct research on the effect of different LLM designs on users. The more models are designed to seem like living beings, or like persons, the less agency users have to decide how they relate to them. The effect is a growing generation of people who make chatbots their most important confidantes, rendering them at the same time more lonely and more dependent on the company’s product.

1258

Rob Horning @robhorning.bsky.social · 09/07/2026

reading reduced to "skill nostalgia" can become part of the political project of deliteracy; this is why "close reading" needs to be taught as a political practice rather than an exercise in aesthetic appreciation and self-soothing

1199

Rob Horning @robhorning.bsky.social · 09/07/2026

post-literacy fears can also be seen as "skill nostalgia" and mourning for the bourgeois public sphere; wouldn't be surprised to see tech ideologues argue both that (1) reading is reactionary and (2) reading is an ennobling hobby aeon.co/essays/the-r...

When earlier periods of skilled practice are treated as more fully human, the social order that produced them can easily appear better than what replaced it. The past, then, seems superior to the present. This is why such arguments risk becoming reactionary. When we mourn the uninvention of skills, we may also – unless we are very careful – end up mourning the loss of the social structures that surround them. That slippage is the core mechanism by which skill nostalgia feeds into a longing for a bygone world.

6249

Rob Horning @robhorning.bsky.social · 06/07/2026

"prediction" markets incentivize profiteers to find weak spots in the fabric of reality and reweave it for easy money — a kind of reality arbitrage, at the expense of our collective faith in a shared world www.bloomberg.com/opinion/news...

Prediction markets are “truth machines,” in the sense that they give people incentives to figure out what is true and bet on it. But they are also, obviously and increasingly, falsehood machines, in the sense that they give people incentives to make bets and then distort reality to make those bets pay off. Because of prediction markets, it is now lucrative to figure out what song will be the most popular, but it is even more lucrative to make a different song look most popular. Prediction markets can make it harder to know what is true.

14715

Reposted by Rob Horning

Rachel Coldicutt @rachelcoldicutt.bsky.social · 22/06/2026

One of the really difficult lessons of the last 100 years is that convenience - drawing on mass market capabilities to do things faster - is isolating and individualising. We become one in relation to the market rather than one in relation to other people. GenAI is the peak expression of that.

212642

Rob Horning @robhorning.bsky.social · 17/06/2026

"why is it pleasurable to know things?" comes to seem like a meaningless question when knowledge is reduced to information retrieval www.techpolicy.press/ais-catastro...

They used to define themselves by their passions, spending years reading, practicing and mastering their fields in order to answer difficult questions. When AI can produce competent analysis in seconds, learning begins to feel unnecessary.

2110

Rob Horning @robhorning.bsky.social · 17/06/2026

the "Eliza effect" is reminiscent of role-playing games— software providing a structure for an emotional attachment to a simulation, more or less for its own sake. LLMs/chatbots make more sense as gaming engines than as "intelligence"

1244

Rob Horning @robhorning.bsky.social · 15/06/2026

Think “slop” is part of the long-standing project of producing consumers who can’t experience satisfaction—no contentment, just more content

1285

Reposted by Rob Horning

Eryk Salvaggio @eryk.bsky.social · 10/06/2026

We often desire the symbol more than what it represents: people go into debt because a new car makes them feel like they’ve “made it.” Generated text is a symbol of text in this sense, a cognitive debt that allows people to “think they have thought” about it when they have moved away from thinking.

29123

Rob Horning @robhorning.bsky.social · 01/06/2026

"AI" means no more good-faith assumptions of a writer's investment and attention on the basis of their text; it imposes a collective and metastasizing burden of paranoia on everyone mattdpearce.substack.com/p/ai-imposes...

Artificial intelligence is an anti-community technology whose principle civic innovation is to reduce social trust and engender a greater sense of unreality in our lives.

519258

Rob Horning @robhorning.bsky.social · 22/05/2026

getting other people to respect you as a "real person" doesn't take place at the level of a style transfer; it invariably involves reciprocal attention and moments of intersubjectivity kneelingbus.substack.com/p/lo-fi-beat...

The problem for the makers of lo-fi imagery is that this existential crisis, if it is indeed what motivates them, can’t be resolved at the level of aesthetics. AI can emulate most legible aesthetic styles easily enough; embedding any sort of recognizable pattern in digital content seems like one of the worst ways to prove your humanity,

151

Reposted by Rob Horning

Josh Shepperd @joshshepperd.bsky.social · 22/05/2026

The most important upcoming scholarly work on AI isn't going to be on how processes become automated, but how subjects (we) begin to lose autonomy, and reflect the regressions of automation in our habits and thought.

918846

Rob Horning @robhorning.bsky.social · 22/05/2026

from John Abromeit's appendix to this collection of Kracauer's writings on propaganda—perhaps useful to understand AI (like local news and right-wing cable news) as an addictive form of terror meant to make cognitive surrender feel like winning cup.columbia.edu/book/selecte...

As we have seen, however,
Kracauer also recognizes that authority— even in a fascist society— does not rest
on violence alone, which explains the crucial role of spectacle and welfare in cre
ating
a “mass addiction” to totalitarian propaganda. Terror also aims to have the
indirect effect of convincing individuals to
willingly
abdicate their own ability
to judge reality and to accept instead the fabricated “reality” of totalitarian pro
paganda.
Kracauer underscores the theatrical nature of fascist spectacle, which
encourages the masses to accept fiction for reality and teaches them that “the
play is more than a play, and madness is the norm.”

2139

Rob Horning @robhorning.bsky.social · 19/05/2026

"artificial intelligence" is, by definition, the negation of human agency, connection, and community

060

Rob Horning @robhorning.bsky.social · 19/05/2026

from Sigfried Kracauer's "Totalitarian Propaganda": propaganda usually seen as being in service of some other aim, but ultimately it becomes an end in itself, a manifestation of power for its own sake

National Socialist agitators do not entice with intellectual proofs and arguments,
but rather seduce with the “primordial energy of rhythm.”46 In fact,
rhythmic formations gain the upper hand as, and to the extent that, the swarm
of previous promises loses its value. Shocks are orchestrated in random material
to prevent any respite from the tension; waves of propaganda are sent out and, as
required, either build steadily or swiftly ebb away; and, in all cases, care is taken
to coordinate tempi to good effect. These occurrences are lent propagandist significance
by the rhythmic quality of the course they take: the skillful sequencing
of quiet and action, of push and pull. If the proclamation of content, of substantive
and attractive goals, was once decisive, now the appeal to the psycho- physical
state associated with the pseudo- reality is predominantly through form— the
form in which the propaganda accomplishes the requisite acts and resolutions.
The further this process advances, the more the nihilism with which the propaganda
handles all convictions reveals itself, of course, and the more undisguised
is the eruption of the will to power, whose product is propaganda. Still, the “primordial
energy of rhythm” could hardly provide lasting compensation for the
destruction of the contents of propaganda if the propaganda didn’t also succeed
in making people forget about its own vacuity by stamping the now transparent
will to power as the content itself

1115

Reposted by Rob Horning

👁️ @nathanjurgenson.com · 18/05/2026

felt like this too neatly linked the feeling of fakeness to the recent pivot to short form video. all metricized and algorithmic media is deceptive. like, what's the "real" way to post content in a system that links attention to profit?

1115

Rob Horning @robhorning.bsky.social · 18/05/2026

it makes more sense to me to say there is no such thing as a "real share" or a "real fan" than to complain about fake ones. who are these "real" people who supposedly are immune to marketing and uninterested in cultural participation? www.garbageday.email/p/everything...

But the entire internet — well, the four biggest apps we use to navigate the internet — has, quite suddenly, morphed into a horrible new version of television, seemingly programmed entirely by low-paid worker-users in the global south. Fake fans and fake shares abound, as every social media platform starts to devolve into some bleary combination of 9GAG and WorldstarHipHop.

130

Rob Horning @robhorning.bsky.social · 13/05/2026

another (and probably less useful) way of putting this is that slop marks the "real subsumption" of content to algorithmic platforms and the capital that organizes them

1276

Rob Horning @robhorning.bsky.social · 13/05/2026

one of my recent newsletter posts has been turned into a proper article parole.cc/compendiums/...

parole.cc

Evidence as content

031

Rob Horning @robhorning.bsky.social · 05/05/2026

it's as though the main use case of "AI" in education is to propagate low-effort hoaxes that undermine the Enlightenment project altogether

0111

Rob Horning @robhorning.bsky.social · 29/04/2026

they never were in the first place journals.sagepub.com/doi/10.1177/...

Today’s social media platforms are no longer centered around sociable activities.

2255

Rob Horning @robhorning.bsky.social · 29/04/2026

Suggests that generated video needs to be mixed with “real”video to thrive; its value is parasitic, dependent on undermining credibility of nongenerated material

1327

Rob Horning @robhorning.bsky.social · 20/04/2026

clips are like AI summaries in that they seem to promise a way to master content overload, but achieve this by making people passive consumers of an ever more superficial stream of disconnected fragments

0111

Rob Horning @robhorning.bsky.social · 20/04/2026

popularity of "clipping" suggests people like consuming highlights, even of (or mostly of) things they wouldn't care about in its full form www.garbageday.email/p/what-happe...

In an essay on X last week, Ed Elson, host of the Prof G Markets podcast, defined the “clip” as something distinct from “short-form content,” writing, “They are specifically snippets of long-form content — things like podcasts, livestreams, and TV shows — a byproduct of something else. In other words, every clip is short-form content, but not all short-form content is a clip.” He argued in a followup post that, counterintuitively, even though a clip, according to his definition, is a slice of a larger product, they have become the main product for the majority of audiences. And the entertainment industry has noticed.

190

Rob Horning @robhorning.bsky.social · 17/04/2026

sounds a bit like Kittler's "discourse network": "the network of technologies and institutions that allow a given culture to select, store, and produce relevant data"

1103

Rob Horning @robhorning.bsky.social · 15/04/2026

these sorts of fakes also shape expectations of what events should look like, and reinforce the expectation that all events should be reducible to images

1143

Rob Horning @robhorning.bsky.social · 15/04/2026

"vibe-living" is good Orwellian term for when people have been made too anxious to make any decisions for themselves about anything www.nytimes.com/2026/04/15/o...

“Everyone Is Using A.I. for Everything” nowadays, a.k.a. vibe-living, and if you don’t, you’re a misfit outsider who should be stoned to death in the town square to prevent contagion, and then A.I. should resurrect you virtually from your data so you can be stoned to death in the virtual town square, for infinity.

0247

Rob Horning @robhorning.bsky.social · 14/04/2026

seems that "AI" makes "you can just do things" and "you can't do anything" into the same thing, as if they are supposed to be the same goal www.nytimes.com/2026/04/02/m...

It’s exactly this dynamic that seems to preoccupy some of our most obsessively agentic individuals. There’s a common prediction among those hoping to outstrip their ovine, mimetic peers: The A.I. models, they say, will eventually supply all of the effort, training and expertise that have historically stood between humans and our ability to simply make things happen. Once all those annoyances are pushed out of the way, the only thing separating the world’s winners from its losers will be the sheer motivation to act, the raw ambition to do a thing — pure, bold agency. You run across this notion all sorts of places, sometimes as a frantically held belief and sometimes as a sales pitch. A sufficiently agentic layperson, someone argues on X, could sit down with an A.I. and produce the same work as a Ph.D student. “In the age of A.I., becoming a super agentic individual is a superpower”; “Being a highly agentic individual is going to be so important with the rise of A.G.I.”; “In the A.I. era, become an ‘agentic individual’ to thrive! Discover how work evolves with A.I. agents. Don’t get left behind!” All that will matter is your own superior drive and will; every other aspect of achievement can be handled, as if by magic, in the guts of some colossal data center.

1104

Rob Horning @robhorning.bsky.social · 10/04/2026

memes are like fashion trends that people "wear" in media by posting; decoding "what they mean" beyond that verges on folly, like asking what different skirt lengths "really mean"

040

Rob Horning @robhorning.bsky.social · 10/04/2026

this conclusion likens social media meme-making with AI paper-clip maximizing; social media is an engine that turns everything into memes the way AI would catastrophically overoptimize on a pointless goal

With so much of our world being so rapidly reshaped for such stupid reasons, it’s hard not to feel as if we’re living in one of the canonical A.I. doom scenarios without even noticing it.

140