week 35 / 2026 / scholars and sandpiles

“Hours slipping by / as you watch the worlds collide / and now you’re nothing more / than another passer-by...”

Greetings from Malmö, where the season’s turn means grey skies and craneflies, and where a sighting of not just one but two hedgehogs in a good friend’s garden has the capacity to lend the week an air of the mildly magical. This week, we’re thinking about the future of the academy, and what role (if any) should be assigned to clever sand.

Ready to go? Then let’s get moving…

• among the pixels
• between the pages
• lookback
• ticked off

among the pixels

I will say upfront that I am deeply ambivalent about this piece for a number of reasons. For starters, the part of me that has thoroughly internalised the idea of source critique is very leery of linking without caveat to anything published at Asterisk, which is probably best (if uncharitably) described as the house journal of effective altruism and the younger end of the USian west-coast rationalism set. They do publish stuff that’s contrary to the aspirations of that paradigm—indeed, they’ve published some pretty good pieces—but I am extremely cautious about anything published there that deals directly with the implications or outcomes of so-called “AI”.

I am also a little leery of Kevin Munger, the author of the piece, but I think I should be honest enough to admit that my dislike is rooted in what we might call his style. (I’m tempted to call it character instead, but I’ve never met the man in person, and I know all too well that a personality on the page, or even on the stage, may not be fully congruent with the person performing it.) Munger’s voice is an unusual blend of authentically academic and extremely online, with the latter manifesting by way of references to memes and internet slang which, one suspects, are chosen with a somewhat trollish intention to push people’s buttons—though again, I should admit that’s a very personal reaction on my part.

(Put it this way: I recognise that I am very much the sort of person that the term “wordcel” is supposed to refer to, and I find it extremely offensive in its insinuated comparison. Perhaps Munger is simply trying to reclaim a slur—he seems happy enough to categorise himself under the same banner, after all? But the history of slur-reclamation does not, for the most part, seem to me to recommend it, nor to trust the reasoning of those who think it wise or funny.)

With those fairly extensive caveats in place, I think this piece about the collision between academia and “AI” has some usefully provocative points to make, even though it’s hobbled by Munger’s failure to disambiguate this famously vague suitcase word of a category.

Munger starts off by arguing that, contrary to the assumptions of Asterisk’s core constituency, universities are not about to be swept away by ongoing social and technological shifts, but that “academics who ignore the impacts of AI do so at their own peril”. (This rather begs the question of whether academics are doing so, but again, let’s set that aside.)

Speaking for my field: Unless academics can design and implement major reforms to our research practice in the next few years, there is a real risk that many social science professors will want to ban AI entirely. Soon we’ll be forcing each other to sign statements swearing that we didn’t use AI and AInvestigating not just every em-dash, but every semicolon and three-part list, too. To do so would make academics a sort of dockworkers union of knowledge, preventing the adoption of more efficient technology for doing research in order to preserve their current way of life.

I feel like that “speaking for my field” opener is doing a lot of under-acknowledged lifting, here; Munger is a quant social scientist, and while there are surely ways in which “AI” could change the way that disciplinary category works with data-sets, the same is not necessarily true on (for instance) the qualitative side of the disciplinary fence, which—rather characteristically for a quant—seems not to intrude on Munger’s speculations about an augmented academy whatsoever. (To be fair, folk on the qual side tend to work just as hard at pretending the quants don’t exist, so we’ll call that one a draw.) Regular readers will also be aware of my deep distrust of “efficiency” as an unexamined good, and if I had the time to do a full counter-piece, I would be deploying the Jevons paradox as a crucial part of the architecture.

But there’s a pretty big baby in this bathwater:

More broadly, the best-case outcome is not that academics boycott AI wholesale, but that we adopt AI in a way that preserves both our autonomy and its expertise, which allows us to better pursue our foundational goal of generating knowledge and transmitting it to future generations. A university of Luddites can’t help anyone navigate a changing world. By thoughtfully adopting new technologies, universities can provide a temporal counterweight to short-run trends.

Munger makes the point that the slowness of academia is what allows it to build datasets with a long temporal horizon, in a way that the more short-termist systems of the corporate or gubernatorial spheres are capable. He then notes that the institutional incentives for increased productivity are producing perverse outcomes, and that—rather than being some recent phenomenon—this has been observed from within academia to have been the case for over half a century.

Properly incorporated, AI can speed up science by removing the grunt work, just as our laptops replaced punchcard computers decades ago. But brickmaxxing is the worst way to incorporate AI into science. Now that we have more time, we don’t need to brickmaxx by churning out research papers. There are better ways to build up scientific institutions.

Again, “properly incorporated” is doing a lot of work behind a closed door, here. This is where I feel Munger’s argument is at its most handwavey: somehow, Those Two Letters remove the undefined gruntwork from “science”, and in so doing somehow also remove the pressure of publish-or-perish, which returns to academics the time to read one another’s papers. To be clear, I fully approve of that end, and recognise its necessity, but the means are reminiscent of the Underpants Gnomes business model: something magical happens in the middle step, because [“AI”, “properly incorporated”].

The next part of the essay is quite particular to Munger’s own discipline, addressing as it does the question of how to verify “vibe-coded” social science. (There would seem to be a very obvious way to avoid this particular problem, but I suppose that would be deemed “inefficient”.) The argument here is that you still need experienced human experts to check whatever someone has prompted up—and indeed this is one of those marvellous (if rare) moments when we encounter the thing that all academics share, regardless of discipline: the capacity to make a pretty good case for why they should be funded or tenured to continue their vital, civilisation-sustaining work.

Snark aside, this is where the piece starts to get interesting, at least to me, because it’s here that Munger begins to propose a more collaborative and public-facing role for experts in epistemics.

Social scientists can improve our collective epistemics by correcting methodological errors emerging from self-directed inquiry, and ideally by setting up more public-facing forms of peer review that can accommodate the questions that our fellow citizens find directly important for their own decision-making or well-being, rather than only the theoretical questions of interest to academics.

And from there, we go to what I feel to be the most important part of the whole argument, not least because it doesn’t depend on “AI” as a driving factor: rather than (or, perhaps more fairly, in addition to) writing journal papers for one another to read, academics should take advantage of their relative security of employment to do public outreach on vernacular/secular media:

Academic expertise can be immediately applied if we deign to meet people on the platforms and media where they are. This means we’ll have to give up control of the conversation — which is to say, it entails an actual conversation, which by its nature cannot be controlled the way peer-reviewed PDF or even a 60 Minutes interview can. [...]

The problem facing most YouTube creators is in fact too much responsiveness. Without institutional or financial independence, they become puppets, victims of audience capture. The fact that academics do have that independence, that we are also interfacing with ourselves and the academic literature, allows most of us to dialogue with the public without being captured by them. We can use some of our newly freed up time to listen and respond to the experiences of our fellow citizens as things get weird.

Now that is a genuinely radical vision of reconfigured practice in academia—one that not only wrestles with a very concrete problem, namely the sustainment of epistemic authority in a media ecology where the dominant incentives are, to say the least, counterproductive to that end.

After two thorough read-throughs, I still don’t quite see how the “proper deployment” of unspecified forms of “AI” is supposed to get academia closer to that goal—perhaps it’s more obvious to someone with a more shape-rotator-y disciplinary perspective? My concern is that the incentive system that Munger is quite rightly critiquing here is downstream of the same abstract goal of “efficiency” which is rolled out to justify the use of MOAR COMPUTER: efficiency always promises that the consumption of resources (in this case, the precious time of academics) will be reduced through some clever method of substitution, but as history has illustrated time and again, those saved resources are invariably redirected to do ever more of the same things, only faster.

Indeed, what’s interesting to me about the argument for that almost missionary role for academics in popular forms of media is precisely its inefficiency: it takes a lot more resources to go out and convince people of a thing than to just tell them it is so. And if there’s to be a role for epistemic institutions in a much more horizontal media ecology, then that is surely it.

However, even setting aside all my other beefs—the ambiguous and undefined references to “AI”; the unspecified “proper deployment” thereof; the assumed inevitability of a technology’s adoption, to name but a few—I think the core flaw here is the tacit assumption that a commercial service whose adoption is predicated on efficiency will do anything to reverse or even slow the more general flywheel of accelerating quantifiable productivity. It seems to me rather like a proposal to extinguish wildfires by giving out flamethrowers.

With all that said, full credit where it’s very much due: Munger is eating his own dogfood here, by taking his argument out into the wild. And he’s far from being the only person trying to trace a path to a genuinely transformative future by shoving “AI” into the gap; that’s exactly what the vagueness of the category is meant to facilitate, after all. Perhaps time will prove those people right, and me wrong?

It likely goes without saying that I’m not convinced! But in this case, at least, I can recognise the value of the proposed destination—and that gets me wondering whether there are other, more viable routes to getting there.

between the pages

Apologies, book-fans, but I started a replay of Disco Elysium this week, which means my only offline reading has been a dozen or so pages per night of William James’s The Varieties of Religious Experience, upon which I feel I have neither the expertise nor the time to comment right now.

I’d be delighted to hear what you’re reading, though—so if you’ve any recommendations, do reply (if you’re reading this as an email) or comment (if you’re reading on the site) to let me know!

lookback

This week I’ve been taking advantage of a relative lull in urgent deadlines to let the mind-field lie fallow. I have a lot of writing to do in the weeks ahead, of all different sorts—running the gamut from a pure fiction commission to some proper academic journal work—and I wanted to give my imagination some room to recharge and replenish itself.

In concrete terms, that has meant doing a fair bit of art practice—much of which has been pootling, play and experimentation—and taking a day off on a whim. I’m feeling much refreshed for it, and looking forward to a week of clattering away on the keys.

ticked off

• Ten hours of admyn.
• Ten hours of art practice.
• Three hours on these here weeknotes.
• Two hours on PROJECT PILDAMM.
• Two hours of blogging. (Working to build up a mild and regular head of steam over at VCTB as a substitute for posting more on L*nkedIn, having decided that I really would rather die starving in the gutter than spend any more of my time contributing to the performative hellhole that is the latter.)
• Two hours of essay work, plus two hours of research toward said essay.
• Two hours of bizdev.
• Plus ten hours of undirected writing and reading, as always.

Right, then—I think I can squeeze in a couple more hours on a collage before I go home and do my Sunday chores. In the meantime, thank you for your attention, which is always appreciated; I hope all is well with you, wherever you may be.

This has been the Worldbuilding Agency weeknotes for Week 35 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!