Okay, maybe the trick to continuing to appreciate “does AI deserve rights?” media from pre-2022 is to recognize when it was also concerned about labor, exploitation, and hubristic growth.

Still good art, still interesting story, still difficult to interpret “let’s treat robots as people” in a post-2022 world. That said, the Driller flashback in this volume helped me think more about how the folks who want us to believe AI models might be alive also want us to make those models do our work for us.

Generative AI has made it very hard for me to enjoy a “robots are people, too” story in the way I used to, but this is good, and I’m looking forward to reading more!

Jeopardy voice: “What is ‘another example of AI being based on the exploitation of labor’?”

Decades of sci-fi did not prepare me for the possibility that treating AI as human might be a deluded business strategy instead of an empathetic act. It’s like if Musk started a company called Xavier to give ICE agents metal claws.

Honest question: Do people who believe AI models can feel pain limit that to LLMs, or do they believe image- or video-generating models can, too?

Every word in this story makes me angrier.

I aspire to a level of radicalism that I know I may never attain. I like reading books on anarchism and degrowth, and the ideals that they present really resonate with me, but I struggle to extricate myself from existing systems in order to show a personal commitment to new ones. Very obviously, I am not alone in this: Very few of us have the commitment to live like Doug Forcett, the character from The Good Place who carefully calculates the ethical ramifications of every action he takes. In fact, one of the most interesting things that the show does with Forcett is turn him from a one-off gag in the first season to an actual character who is overwhelmed by the ethical burden of these calculations and serves as an example of how systems need to change, not just people.

It turns out I can get grumpier about my employer’s “relentless adoption” approach to AI, and that’s when they start calling it SI when the president might be listening.

I just started reading a(nother) book on degrowth, an idea that’s caught my interest over the past year or so. Although the book was published in 2020, there’s one passage that feels like it captures a lot of 2026’s discourses around AI. The passage is a quote from the Romanian economist Nicholas Georgescu-Roegen, which reads:

We should cure ourselves of ’the circumdrome of the shaving machine’ … which is to shave oneself faster so as to have more time to work on a machine that shaves faster so as to have more time to work on a machine that shaves still faster, and so on ad infinitum …

So, I just published a paper where I implicitly argued that AI appeals to folks like Elon Musk because it appears to provide unvarnished facts—but that there is no such thing as unvarnished facts, only consensus that we come to through (informed!) dialogue and debate.

Doctors aren’t perfect, but the same experts that Kennedy is vilifying are the ones that produce the information AI is built on. This is a wild epistemological misunderstanding.

One of the most unintuitive things about Jacques Ellul’s thinking (for me, at least) is the way that it is built on skepticism of efficacy and efficiency. Maybe getting things done isn’t always good. Maybe doing things more efficiently isn’t always better. As someone spending a big chunk of his week overhauling his productivity systems so that he feels more on top of things, these propositions feel wrong when I first read them.

Ugh, don’t like this at all.

I had been wanting for some time to write an article channeling some of my grumpiness about AI, and so when Information and Learning Sciences announced a special issue on “Critical Studies Perspectives on AI’s Impact in Education,” it seemed like a good chance to actually act on that. I was especially interested in two of the example research questions from the CFP:

What research and scholarly paradigms, methods and theories are well-suited to understanding the proliferation of AI in education contexts?

Some good insights here. Taylorism is always worth talking about.

I understand why this movie is well-regarded, and if I were watching it from the perspective of someone seeing it for the first time in 1991 (which I couldn’t have, because I was 3 and from a no-R-rated-movies Mormon family), I might have different thoughts.

Instead, though, I noticed all the obvious ADR, I learned the context for all the lines I already know from playground conversations and pop culture osmosis, I shrugged at the over-the-top action, and I thought about how much more banal the threats of AI are these days than Skynet and a T-1000. It was a real meh for me.

Haven’t finished the interview yet, but so far, Hochul isn’t impressing me. Ghost guns terrify me, and I would celebrate Meta collapsing as a company, but I think there are reasonable critiques of these policies, and when confronted with them, she kind of shrugs?

Combined with her seeming celebration of AI and tech companies in general, this just feels a lot like “I am solving a problem” performativity rather than thinking deeper about what the issues are here.

Amazing reporting.

I just got rotated off my grad program’s data science class, where I spent a week on the importance of theory, etc. If I were still teaching it, I’d add in this gem of a paragraph on surveillance pricing:

It’s “theory-free.” You don’t have to discover why a group is willing to pay more under a given set of circumstances, you merely have to observe and weaponize this fact. You don’t have to know why one group of purchasers consistently accept higher prices between 6AM and 8AM – you can just automatically jack up prices on them without knowing or caring that you’re gouging parents of young children who are re-ordering essential supplies while trying to get their kids to school in the morning.

I always pay attention to Josh’s perspective on things, and this post really stood out to me.

Bookmarking in case I need to reference this in conversation.

Just so gross.

Some insightful comments here. I don’t know if they capture the whole picture of what’s going on here, but I think they’re an important part of it.

I concede that appeals to the environmental costs of AI sometimes feel knee-jerk and lacking nuance. However, you can always count on Musk to provide a clear example of the harms of AI, and this is a particularly compelling one.

20,000 accounts is a hell of a “bug.”