Tessera
Folgen

Dez Miller

@dezmiller.bsky.social

urban rivers | DH and Critical AI Studies | Post45 lit and film Dezmiller.com

179 Posts419 Follower333 Folgt
Dez Miller@dezmiller.bsky.social · last wk.

"Nolan’s Odyssey lacks many of the elements that make the poem great...The writing is abysmal. None of the characters has convincing motivation for their actions or words. There are no sex scenes, and all the food looks horrible." 😂 (she ends by saying she's glad Nolan is getting ppl to read)

Emily Wilson · An Uncomplicated Man

As they say at the awards shows, I was humbled to learn that Christopher Nolan has read at least the first line of my...

lrb.co.uk

Alondra Nelson@alondra.bsky.social · 2w ago

Students from all 50 states just drafted an #AIBillofRights. They named cheating, bias, privacy mental health, environmental costs and more. The civic grammar of AI rights is traveling and we're fortunate that young people are leading the way and writing it. www.the74million.org/article/stud...

"High school students, steeped for nearly the past four years in generative artificial intelligence, face this reality daily, attending class largely without a set of agreed-upon guidelines. Over the weekend, a small group of students came together to hammer out a few basic principles for the technology, something akin to an AI Bill of Rights. Sponsored by a group that includes an MIT research initiative and a national superintendents’ association, 100 students from all 50 states gathered here to debate a first-of-its-kind set of recommendations, to be shared with district leaders nationwide as AI proliferates rapidly in the lives of both students and teachers."
Alondra Nelson@alondra.bsky.social · 3mo ago

AI companies are writing their own constitutions. Meanwhile, a shared vocabulary for contesting algorithmic power has been traveling across red and blue states, from legislatures to civil society. New from me in @science.org: "A Civic Grammar for AI Rights" www.science.org/doi/10.1126/... 🧵

Anna Kornbluh@annakornbluh.bsky.social · 2w ago

the best day for the presidents of every R1 to hold a press conference screaming that higher ed is our second largest export and destroying it is economic sabotage and soft power detonation was Jan 21st 2025. the next best day is tomorrow.

Anna Kornbluh@annakornbluh.bsky.social · 2w ago

a bad thing is running a phd program with brilliant international students who must reckon with a volatile xenophobic saboteur government in fresh ways every day, no thank you, 0/10

Jordan S. Carroll 🌹@jordanscarroll.bsky.social · last mo.

OK, let's imagine a world where the humanities never goes into crisis in the 2010s. let's say a multi-billionaire tech dies and his will lists "the humanities" as his only beneficiary. what does your field look like? Here's what I think English would be. /1

Matthew Kirschenbaum@mkirschenbaum.bsky.social · last mo.

You can split the hair finer and finer—“I don’t mean all AI”—“I don’t mean all ML/AI”—“I don’t mean all generative AI”—but telling academic colleagues let alone the public at large there is no responsible use case for AI is not going to be a message that gets a “critical” position taken seriously.

Christina Ayiotis@christinaayiotis.bsky.social · last mo.

“Artificial intelligence programs can spot patterns in electrocardiograms that humans miss. Now, one program is going to be widely available — for free — to doctors.” www.nytimes.com/2026/06/22/h... @nytimes.com

Laura K. Nelson@lauraknelson.bsky.social · 2mo ago

Genuine question: Do we think this linked framing holds up after the explosion of AI agents? "Like catalogs and the internet, large models are part of a long history of cultural and social technologies." Are LLMs (now) really in the same category as card catalogues? www.science.org/doi/10.1126/...

Large AI models are cultural and social technologies

Implications draw on the history of transformative information systems from the past

science.org

Julia Carrie Wong@joolia.bsky.social · 2mo ago

humans have known how to construct a shoulder seam for thousands of years, but with AI, we are unlocking untold new ways to fuck it up

Citizen.Coping@propcazhpm.bsky.social · 2mo ago

Teams have to wash and steam #WorldCup shirts to remove shoulder bulges. "Nike boasted about the use of 'computational design'… driven by performance data that incorporated elements of AI to work alongside the company’s designers as they crafted the kits." How could they even ship these?! Shame.

Shannon Mattern@shannonmattern.bsky.social · 2mo ago

"Replica of a Chip: The Weaving Technology of Marilou Schultz marks the first survey of acclaimed Navajo/Diné weaver + math educator Marilou Schultz... The exhibition positions Schultz as an innovator whose work... has influenced the practices of art, Navajo weaving, + computer architecture..."

Replica of a Chip: The Weaving Technology of Marilou Schultz - CCS Bard

Opening Reception, Saturday, June 27, 2pm - 5pm Limited free seating is available on a roundtrip chartered bus from New York City for the June 27 opening. Reservations are required and can be made on ...

ccs.bard.edu

Ryan Heuser@ryanheuser.com · 2mo ago

Surprising: frontier models (Claude, ChatGPT, Deepseek V4) produce the most predictable text of any local AI model or human text I've ever tested. Not surprising: Finnegans Wake is off the charts, by far least predictable—Shannon & Lydia Liu proved right—and Hemingway the most predictable human.

Horizontal boxplot comparing information density (BLT 1B bits/char) across 27 text sources, ordered from most compressed (top) to most information-dense (bottom). Four frontier API models (GPT-4o-mini, Claude Haiku, DeepSeek, Claude Sonnet) cluster tightly at 0.85-0.90 bits/char, below Shannon's English rate of 1.0 bits/char marked by a dashed vertical line. Seven local aligned models (red) span a wide range from DeepSeek-7b aligned (0.94) through OLMo aligned (1.43), with most falling between 1.0 and 1.3. Seven base models (green) occupy 1.04-1.50, overlapping substantially with human text. Eight human text sources (blue) range from Hemingway (1.12) and abstracts (1.20) through dream reports (1.24), waking journals (1.26), and Basic English stories (1.39) up to C20 fiction (1.50) and Joyce (2.30). OLMo aligned is a notable outlier among aligned models, with higher information density than its own base model. DeepSeek-7b base (1.04) is unusually low for a base model, sitting near Shannon's threshold alongside Hemingway.Lydia Liu, The Freudian Robot, p. 37: "...interesting questions that his experimental work raises for us is: how does a stochastic view of writing correlate to the received theories of language, literature, and modernism on the one hand and to psychoanalytical speculations about the unconscious on the other? This question is pertinent to our inquiry because many of the earlier modernist literary and psychoanalytical experiments on language, automatic writing, and thought-reading had anticipated Shannon’s Printed English and his “mind-reading machine” in numerous ways. For instance, Shannon cites James Joyce’s Finnegans Wake as one of the texts exemplifying the lower threshold of redundancy and higher entropy rate in his stochastic model of Printed English. What makes entropy and its possible linkage with Freud’s Todestrieb (death drive) such an interesting problem for the study of digital media is the ways in which certain ideas migrated into psychoanalysis first and then got into information theory. Furthermore, Freud’s work and psychoanalysis in general may suggest some interesting clues as to the shared theoretical impulses or implicit exchanges among information theory, cybernetics, and modernist literature. Spanning across these moments of broad intellectual confluences is the techne of the unconscious that continually articulates itself to digital writing, machine, and social engineering. We turn next to the invention of Printed English by Shannon and its implications for a theory of digital writing."
Ryan Heuser@ryanheuser.com · 3mo ago

Shannon measured the information rate of English at ~1 bit per character. According to a byte-level LLM measuring next-character predictability in LLM & human text (diaries, abstracts, dreams, fiction), aligned models produce sub-English information rates & LLM text is more predictable than humans'.

Bar chart comparing information density (BLT bits/char) of AI-generated prose versus human text. Shannon's English rate (1.0 bits/char) shown as dashed red line. Aligned OLMo models (SFT, DPO, RLVR) fall below the line at 0.89–0.99 bits/char. The base model sits just above at 1.14. All human text types are higher: waking reports (1.24), abstracts (1.28), dreams (1.32), and fiction (1.49). Alignment compresses model output below the information density of all measured human writing.
Jonathan Basile@jonothingeb.bsky.social · 2mo ago

These are the more practical conclusions: I argue that we arrive at a more robust critique of 'AI', and a more thorough understanding of artistic and linguistic production (from human and machine), when it is not based on intentionalist theories of meaning

The LLM is no more a self-conscious subject than we are, but this removes one commonplace stopgap for critiques of ‘AI’. Its statements can no longer be claimed as meaningless, or its artworks as devoid of aesthetic quality, simply because it lacks the intentions of a subject. Our own language has never depended on our intentions for its meaning (as we can see, machine writing remains legible—so much so that at times it is difficult to distinguish or prove its provenance—even if we may tend to read it differently than human writing, when we know or think we know its origin). A structuralist view might conclude that language, as a system unto itself, not only does not depend on such intentions but is the matrix in which they are generated, that all the effects we attribute to our own subjectivity or initiative are effects of language, misrecognized. Derrida was never satisfied by such notions of the ‘death of the author’: Rather than oppose citation or iteration to the noniteration of an event, one ought to construct a differential typology of forms of iteration, assuming that such a project is tenable and can result in an exhaustive program, a question I hold in abeyance here. In such a typology, the category of intention will not disappear; it will have its place, but from that place it will no longer be able to govern the entire scene and system of utterance. In other words, intention does not disappear from the scene once it is deconstructed—it is just that it, like everything else, is written and read. It cannot govern the meaning of a statement (we argue over presumed intentions or can dispute explicit intention-claims, can dissimulate intentions and even be at odds with ourselves over them, and can see further possibilities of interpretation or reading beyond a given intention, etc.). It follows from deconstruction not that intentions and statements of them are meaningless and not that language or writing univocally masters the scene, creating a meaning of and unto itself that forecloses all others, but that intentions are textual. In other words, we know them by that relay described above, understanding our own through language, the world, and the other. They exist for us or an interlocutor or witness only to the extent that they can be inscribed (implicitly or explicitly), which means that they are not the origin of language or action, but rather a participant in the scene of writing. Then, our own ‘intentions’ can be transformed by inserting a novel writing machine into our circuits, and that machine may produce intention-effects just as we do, by displacing those relays, by taking part in the self-exteriority of textuality. That certainly does not mean that we ought to treat the writing of LLMs just as we do writing from other sources (except in the sense that both must be read, one way or another), it only means that the linkage we try to form between text and context is insecure in both cases. Nor do we need to seek out our inner light to dismiss the artistic productions of ‘AI’. While it is often said that some form of genius, sympathy, or suffering sets apart human artistry, we should recognize that whatever self-relation we accomplish through art-making (a) necessarily subjects itself to established and iterable art-forms, genres, etc., even or especially when it attempts to exceed their limits, and (b) first comes to itself in a relay with such formulaic art (life imitates art, etc.). Again, it is what we have in common with writing machines that allows us to read them otherwise. Context plays a nonprogrammable role in how we read anything, including an artwork. The stories we can tell about a work and its author, the history and tradition within which we contextualize it, and the environing scene in which we situate it all play some role in how we interpret and relate to the work. This context does not determine or constitute reading any more than intention does (nor is it a simple given what counts as context); context participates in the scene. So, anyone is free to imagine that the products of ‘AI’ emerge from a thinking and feeling computer consciousness that they then relate to in an identificatory fashion (as they would any other artist—but perhaps not exactly so), or to see such works as the lazy detritus of a human machine-user or prompt writer, worthy of no more of the reader’s effort than was put into their creation. even if we share a disdain for the type of poetic or artistic work thus far created with ‘AI’, we should not try to theorize its deficiency by claiming that all automated or programmatic processes are inimical to creativity. As Derrida often wrote, it is necessary to think ‘the event with the machine’. Invention or the unanticipatable is not only possible as a repetition that changes contexts, but every invention takes place this way. Long before LLMs, we had poets and artists making use of much simpler algorithms, and even before the computer, combinatoric processes were used to bypass the censors of consciousness; moreover, processes of recontextualization (such as found art or poetry) demonstrate the interrelation of invention and iterability. Even if such artistic practices remain matters of taste, I would argue against dismissing them on the grounds that art ought to emanate from the mystical font of pure creativity; if ‘AI’ art disgusts us today, if there is a truly visceral reaction against it, that is perhaps not because of its reliance on machine or iterability (always already the case), nor because of its uniquely bad aesthetic quality (the average and majority of human composition has been similarly slop-like, but may evoke a different reaction); rather, the context and story surrounding this technology is part of our aesthetic experience. That it is dismantling the economic possibilities for human artists and many others, forced upon us by a capitalist system fundamentally hostile to life and nature, serving as the fun-and-innocent front of an ideological campaign building infrastructure and ideological support for ever more totalitarian forms of surveillance, warfare, and governance, all for the benefit of eminently mediocre man-children (too lacking in taste to even appreciate what they are destroying, except for perhaps a barely conscious resentment), this shapes aesthetic experience in a manner that a theory of intentionalism can only obscure.
Jonathan Basile@jonothingeb.bsky.social · 2mo ago

I wrote an essay exploring how structuralism and deconstruction have informed recent theory around 'AI' and 'LLMs' including @leifw.bsky.social's Language Machines: olrsupplement.com/2026/06/01/w...

Anna Kornbluh@annakornbluh.bsky.social · 3mo ago

universities have the brainpower, resources,+ community obligation to develop on-site secure NON PROFIT learning management systems + ed tech overseen by actual educators, not data-extraction censorious surveillance profiteers! VOID THE CONTRACTS theamericanvandal.substack.com/p/mamdani-wi...

Mamdani Win Could Be The First Step Towards Seizing The Means of Knowledge Production

Let CUNY socialize EdTech for all of us.

theamericanvandal.substack.com

Dez Miller@dezmiller.bsky.social · 3mo ago

Ok I may be in a “cat has chosen our family” situation. Young stray cat keeps trying to dart into our home and is basically living on the porch. Never cared for a cat before. Help!

Coalition for Networked Information@cni-org.bsky.social · 3mo ago

The John W. Kluge Center at the Library of Congress is offering Kluge Fellowships in Digital Studies for scholars using digital methods to explore the impact of the digital revolution on society, culture, and international relations. Open to all disciplines—apply by Sept 15. #DigitalScholarship

Kluge Fellowships in Digital Studies | Fellowships | Chairs & Fellowships | The John W. Kluge Center at the Library of Congress | Programs | Library of Congress

About the Fellowship The Kluge Fellowship in Digital Studies provides an opportunity for scholars to utilize digital methods, the Library’s large and varied digital collections and resources,…

loc.gov

Richard Jean So@richardjeanso.bsky.social · 5mo ago

I'm excited to be a co author on this new paper, "Computational Hermeneutics," with a bunch of other great scholars from the humanities + computer science. In it, we lay out concepts for evaluating gen AI's capacity for interpretation esp ambiguity, context, etc. www.frontiersin.org/journals/art...

Frontiers | Computational hermeneutics: evaluating generative AI as a cultural technology

Generative AI (GenAI) systems are increasingly recognized as cultural technologies, yet current evaluation frameworks often treat culture as a variable to be...

frontiersin.org