justine zhang

@tisjune.bsky.social

assistant professor in information, university of michigan. i have big intellectual feelings about language and technology. https://tisjune.github.io/ she/her

In re: to recent news about the Meta data center, Sheila Tahir and I are sharing our zine, "Ghosts in the Cloud: A Southern Gothic Approach to Studying AI Infrastructure" based on a trip to the hypderscaler construction site in spring 2026.

Got a useful question on the thread about student visas as to whether it's contradictory, from an abolitionist perspective, to describe the university as "a sanctuary for knowledge." The answer, for me, is both yes and no. Sharing an elaboration here.🧵

Why Bernie Sanders is wrong - and nationalizing AI is a bad idea "AI is, above all else, a weapon of class war from above. Whether it’s the state that wields it or corporations does little to change that fact."

The Socialist Case Against Nationalizing AI

AI is not a neutral technology whose benefits need to be redistributed. It is a weapon of class war from above, designed to cheapen labor and concentrate power. Nationalizing it does nothing to change...

jacobin.com

A university cannot be a sanctuary for knowledge while serving as an agent of the border. As long as some students can be threatened with deportation, no one's academic freedom is secure. 20/

much "critical" work on AI can be understood as attempts to elaborate an ethical consumerism around AI. that this work systematically exhibits contradictions is unsurprising, as ethical consumerism is a wildly inadequate frame for contending with the violent, hegemonic projects that AI indexes.

at facct this week? wondering if "political economy" is a rich collection of intellectual tendencies and critical sensibilities, or concerned that it's yet another academic branding tactic?curious about the background text of this flyer? join us!! sat 7-9 at co-op BMP!!

political economy of algorithms and computing (a facct 2026 social)
coop bar milton parc sat june 27 7-9pm
black text over a background of grey text, an article titled "class capitalism and computers" from the 13th issue of interrupt, a newsletter published by erstwhile org computer people for peace

something i've been thinking about is how such initiatives are ways for universities to convert political challenges (as seen in efforts at divestment and worker organizing) into speech problems, hollowing out their content while making them easier to manage into submission

Hypervisible @hypervisible.blacksky.app · 2mo ago

“Based on the job descriptions, the new positions will consolidate into one role a variety of free expression–related responsibilities—such as crafting speech policies, training faculty in de-escalation tactics, monitoring protests and coordinating with law enforcement.”

Just finished reading and this is all so very important and spot on. There's value in countering specific claims - such as in my case AI's ability to actually create software not just generate code - but discourse whack-a-mole just flings rocks at the walls while the castle remains standing.

An except from the above paper with the following text in which I've highlighted the first and last paragraphs:

As we’ve outlined in §2, a large body of scholarship has engaged with these discourses as collections of mistaken or absurd claims, to be debunked, corrected, or ridiculed with better scientific techniques, deeper engagements with social scientific theory, and more measured accounts of how the technology really operates. Such strategies can serve as useful ground-clearing for broader efforts, but are limited in their ability to check the power of the AI industry, because they risk drawing attention away from the ideological work these discourses do [29].

In the context of AI capabilities discourses, claims about what AI can do carve social life into discrete skills, capabilities, and use cases, and rerender these things in terms of AI. If the point is to contest the broader systems of exploitation that the project of AI is woven into, then simply offering better assessments of the ability of an AI model to do productive cognitive work leaves intact the reifying logics that continue to alienate people while organizing capital.

Critical scholarship on AI should therefore aim to interrupt the ideological work performed by these claims, dismantling the grounds on which they are sayable rather than simply interrogating their truth value.

my paper "Hyperscalers as logistical regimes" was published last week! Read on for a story about how Silicon Valley adopted ideas from the world of shipping to make (and to some extent, fulfill) promises about scale and the fluidity of the circulation of capital www.tandfonline.com/doi/full/10....

Hyperscalers as logistical regimes: software containers and the circulation of cloud capital

This article analyzes the software container as a device central to the rise of cloud capitalism. Firms like Google pioneered software containers internally as infrastructure for their own producti...

tandfonline.com

if i had to pick one moment to put into that "you bolt awake in the mountains of carthage" meme it would probably be when people started earnestly talking about using AI agents to do stuff

feeling sad and fantasizing about the two semester seminar style course i would teach if i had, like, a completely different job... semester 1 would be focused on language and political economy, often together... semester 2, of course, would be "oops all wittgenstein"