week 12 / 2026: agents and agency
This week’s reading notes take a blog post by Matt Jones as their springboard, and as an invitation to ask the unacknowledged political-economic question lurking beneath even this fairly sane and sober look at the so-called “agentic” revolution. That question is: “do we really need or want this?”
“Now your blood / travels though the veins / of our history…”
ISIS
Greetings from Malmö, where the sky is blue and the air is crisp and by the gods we’re glad to be on the right side of the equinox. Welcome back to WEEKNOTES at Worldbuilding Agency where, yes, I am going to be writing about Those Two Letters, because taking regular glances into the abyss du jour is part of the job description.
Loop the rope around your waist, and strap your goggles on tight…
among the pixels
My springboard this week is a blog post by Matt Jones (ex-Google, ex-BERG, ex-Miro), for two reasons: first of all, a significant number of people sent me the link by various channels to ask me what I thought about it; secondly, Jones is unusual for a software guy, in that he’s actually spent a bunch of time engaging with social theory and the history of science and philosophy.
As a result, the post goes from the programming term-of-art “wall-clock time”, through the various “AI brain-fry” studies of recent weeks, Steve Yegge’s “Gas Town” hypomania, vampirism, E P Thompson’s seminal study of clock-time as the pulse of industrial capitalism, Felix Guattari’s distinction between human and machinic time, and Paul Virilio’s “science of speed”, dromology.
(If you’re unfamiliar with the first three of those references, well, lucky you—like I say, gazing into the abyss is part of the job, and it’s not always the fun part. As for the others, that might seem a lot of pretty deep sociological cuts for a blog post, and I’d be inclined to agree; I’ll return to this point further down.)
Longish story short, Jones compares the increasingly obvious physiological impacts of intense LLM usage to the reported impacts of earlier infrastructural reconfigurations:
[Virilio’s] argument was that the history of civilisation is not primarily a history of wealth or territory but of velocity: who controls the fastest, densest barrage controls the territory. Each new speed technology — the stirrup, the railway, the telegraph, the missile, the fibre-optic cable — reshapes not just logistics but perception itself.
Speed doesn’t just let you move more easily; it changes what you can see, hear, and think. Push acceleration far enough and you get what Virilio called the “aesthetics of disappearance” — things moving too fast to be perceived at all. The landscape seen from a bullet train isn’t a landscape anymore; it’s a blur. The high-frequency trade executed in microseconds isn’t a decision anymore; it’s a reflex of infrastructure.
He then talks about the “bullet hell” genre of video games, to talk about the difference in attentional style and affect between cognitive overload that is taken on voluntarily, as a form of entertainment, and cognitive overload imposed by external circumstances… and from there, we get a Fury Road reference that builds upon and to some extent complicates Yegge’s adoption of that movie for his own naming scheme.
Jones closes by framing all of this as a design problem—which, to be clear, isn’t wrong per se, but (to me, at least) desperately incomplete. I remarked to one of the people that sent me the link that I found it a deeply frustrating post because it walks all the way down a long hallway and then declines to open the big red door at the end of it.
The sign on that door reads “political economy”. While the political aspects of Marx passed their sell-by date a very long time ago, the economic aspects of his thinking are incredibly relevant right now, because so much of Capital was a thinking-through of an upheaval in the mode of production that is at least comparable to the one that we’re supposedly living through now: the rapidity and churn, the blood on the floor, the top-down and heavily-subsidised narrative of inevitability underpinned by a logic of efficiency and growth uber alles.
(As an aside, this is why I have never considered the label “Luddite” to be an insult. Here’s Thomas Pynchon explaining why, nearly three decades ago.)
The question that Jones leaves unanswered is “do we really need this, is this really what we want?” To suggest design fixes for a technological superstack that you’ve just successfully described as vampiric, physiologically damaging and fundamentally corrosive to the social fabric is to accept the underlying logic: that an increase in throughput—which Jones identifies as “the virtue of the agentic [age]”—is worth the sacrifice of reconfiguring the sociotechnical structure. He describes this as a “domestication” of temporally distorting technologies, and asks “whether we can do that faster this time”.
But as regular readers will know, “technology” is not a distinct category; it is not separate from us. (This is the cyborgean condition that Haraway was talking about back in the 1980s.) The domestication Jones suggests is thus in truth a habituation—a conditioning of ourselves to these new super-productive prostheses. And to pose the ultimate question as being whether we can achieve that habituation faster than before is a bitter irony: the logic of Bigger Better Faster More is fully internalised, beyond question or reproach.
Now, I’ve been reading Jones for years; I even met him once, back in the BERG era, though I doubt he remembers it. I don’t think he’s some sort of moral monster. But I do think that he, like all of us, is a product of his own context.
It’s worth noting that many of the people most engaged with and enthusiastic about the supposed “agentic future” are people whose work-lives are largely self-determined and secure. While they often talk in terms of compulsion and irresistibility that always remind me strongly of that period in the late Nineties when a lot of my friends got really into cocaine, these are people for whom their early-adoption and experimentation is a choice.
As such, I think it may be hard for them to fully appreciate the rather different perspective of people for whom technological systems are already the cage of impositions that defines not only their work life but, increasingly, their social lives as well. To use a term popularised by Cory Doctorow—no, not that one—it’s hard to swallow the angels-on-a-pinhead philosophical discourse of centaurs when you yourself are a reverse-centaur.
As I suggested above, Jones actually walks right up to this question in his “bullet hell” section: a cognitive barrage can be thrilling when it’s chosen, but when it’s not, it’s just endless trench warfare. Could that problem be designed away? Hypothetically, perhaps. Will it be designed away? In a world where human beings scrape a living as the squishy components of an Amazon warehouse or deliver morning drinks by moped to knowledge workers too busy to make their own damned coffee, I hope you’ll forgive me for thinking it extremely unlikely.
The “agentic” era is presented as revolutionary, but it is quite the opposite: rather than overturning the logic of the preceding era, it is instead intensifying and clarifying it. If you want to think about the futures of “AI”, or indeed any other supposedly technological upheaval, you have to look at the context as well as the sexy new figure in the foreground.
My skepticism about the whole business is less to do with a refusal to understand that “the genie won’t go back in the bottle”*, and much more to do with my inability to resolve the yawning contradiction between the promise of the “tech” and the fragmenting social structure into which it is being heedlessly rammed. That a well-read and well-intentioned person like Jones can’t seem to walk through that door serves mostly to vindicate my skepticism: assuming these systems can do everything that is claimed for them (which I also doubt), then deploying them at scale will have an effect comparable to throwing a grenade in a fish-pond.
To return to the reference-heaviness of Jones’s post: to his credit, he explains in a coda that it was written with significant input from an LLM, and reflects on the sensation of feeling like “a court jester or the class clown in the seminar” in the process. Again, this feels a little like walking up to the really important part of the “constellation” of ideas and then kind of shrugging it off. In the name of getting the post done quickly, the part of writing that is the thinking-it-through has come to seem like surplus effort: why bother, when an algorithm can just spit out and rewrite material in a convincing facsimile of your own voice in mere minutes of wall-clock time? For me, this feels a bit like winning a marathon by doing it on rollerblades.
To build on Jones’s own metaphor, we’re shown the constellation, and a set of lines that suggest a figure made from those stars, but we do not get the deep story of the deity or mythological being that the figure represents. What’s missing from this piece, and many others like it, is synthesis: the pulling together of ideas into a conclusion that takes you somewhere new, that takes you through the door at the end of the hallway.
The hill I will die on is that an LLM will never be able to do that work. Such a system can assemble tokens into something that convincingly resembles a synthesis, but it cannot think; it can juxtapose, but it cannot combine. As such, Jones’s post is, perhaps accidentally, an illustration of what we stand to lose in the pursuit of the “agentic”—namely our own agency as thinking creatures, willingly traded for greater throughput.
This is the tragedy and hazard of what John Willshire has taken to calling “cognitive debt”. To return to my own metaphor, it’s like you rollerblade most of the marathon, but at some point get distracted and head off in a different direction to the finish line. The important part, as well as the hard part, is set aside, left undone, in favour of the pure thrill of velocity; all speed, no direction.
If you can’t see in that image an absolute crystallisation of the zeitgeist, I don’t know what to tell you.
( * The genie-in-the-bottle metaphor has always amused me, in a gallows-humour sort of way, because it manages to overlook the key feature of that story-form: the genie always goes back in the bottle, usually right after you’ve spaffed away its actually fairly limited powers on poorly-thought-through wish-fulfilment.)
between the pages
I’ve run unusually long in the previous section, even by my own prolix standards, so here I’ll not simply that I’m about a quarter of the way through Thomas Mann’s The Magic Mountain, and enjoying it hugely.
Apologies, literature fans; there will definitely be thoughts to come about the affordances and challenges of the involved or omniscient point-of-view, and its rarity in modern fiction—but not this week!
lookback
Busy week. I was up the cast in Gothenburg on Monday and Tuesday, playing guest lecturer to design postgrads at HDK Valand on the former day and preaching the gospel of worldbuilding as a framework for foresight and strategy to the G-town chapter of Futures Sweden on the latter, as well as squeezing in a few meet-ups and conversations.
Lots of projects are spinning up in the back end of March, which has meant a lot of digital meetings, and the construction of a very tightly-packed schedule for next week. I’m rather behind where I’d have hoped to be by now, as a major power cut on Wednesday morning somehow knocked out the broadband at my studio, and it didn’t come back until Friday morning. This mean I was obliged to do anything requiring bandwidth at home, in the absence of my regular office infrastructure; as discussed above, habituation to one’s technological prostheses is just fine until forces beyond your control snip off a bunch of your metaphorical limbs!
I made an omelet from those broken eggs, though, by shifting my art time to Wednesday, and doing client work and admyn over the weekend. There’ll be more done after these weeknotes go out, too! The flexibility of the freelance life is one of its upsides, but it sometimes comes at a price.
Well, selah—buy the ticket, take the ride, as the man once said.
ticked off
- Seven hours of teaching at HDK Valand.
- Seven hours of admyn.
- Six hours of networking.
- Six hours of art practice.
- Five hours of talk prep and delivery.
- Three hours on these here weeknotes, and two hours editing and pasting up my interview with Joost Vervoort.
- Two hours on PROJECT MUNICIPAL.
- Two hours on PROJECT PROSCENIUM.
- Two hours of planning and development for an online version of my Fiction4Futures course.
- Two hours of academic peer review.
- One hour on PROJECT HORNIMAN.
- And ten hours of undirected writing and reading, as always.
Blimey—got your Virilio-esque velocity, right here, mate.
Right, there’s more to be done on PROJECT PROSCENIUM before I can down tools for today, so I’m gonna tie this off here. I hope all’s well with you, wherever you may be.
This has been the Worldbuilding Agency weeknotes for Week 12 of 2026. Thanks for reading! If you've enjoyed them, it's free to subscribe. If you are already subscribed, please send to a friend!
Comments ()