Join free with your email
Free. No spam, ever — we'll never share your email address and you can opt out at any time. Already a subscriber? Log in
Researcher with public evidence across AI research, Culture, work & education.
“Nearly all aspects of the accounts appear to be AI-generated — from the “person” in the video to automated responses to comments, which in some cases attempt to mimic African American vernacular — and experts warn scams like this are growing every day.”
Don’t let Meta near your kids.
We fucking told you so. We told you and told you and told you. "AI" as an industry has to earn $2 trillion in revenue per year from now to 2029 in order for investments merely to break even. That's simply impossible. www.theverge.com/ai-artificia...
Yes, good. More like this.
arxiv.org/abs/2508.01491
“Sooner or later, elite schools are going to realize that an analog education carries more prestige than an AI-assisted one & will retool their curricula..to reflect that prejudice. This has .. already happened with screens in elementary and secondary education” hedgehogreview…
“Every day, Michael Geoffrey Asia spent eight consecutive hours at his laptop in Kenya staring at porn, annotating what was happening in every frame for an AI data labeling company.” www.404media.co/ai-is-africa...
how to turn the AI stuff off: www.washingtonpost.com/technology/2...
The Commerce Department eased its block on Anthropic's Mythos, but UC Berkeley Risk and Security Lab's Andrew W. Reddie argues Lutnick's letter leaves the hard question unanswered: in an age of cloud-hosted models, what even counts as an "export" when a capability crosses bord…
New plaintiffs in a lawsuit against Elon Musk's SpaceXAI and Stability AI say the companies' AI tools were used to make sexually explicit images of them as children. n.pr/4wu1xkr
A growing assumption in AI safety is that frontier systems will fail unpredictably rather than through coherent misalignment. But as Jennifer Kinne argues, governance may be missing a more dangerous failure mode: models that appear reliable while gradually drifting away from r…
I’ve been wondering for a long time where they got the idea that he was going to send them (the very specific number of) $5000 checks as a “refund from DOGE.” It seemed like a mass hallucination. I’ve just seen screenshots. They send boomers/retirees texts about it. I should h…
We have entered the era of advertisers shitting on AI. I just saw a commercial for Legal Zoom where two folks are trying to set up their will and one of them suggests Legal Zoom, and the other says "we'll just have AI do it." Turns out the AI forgot their 401k, which the human at Legal Zoom caught.
Forever angry that there are some useful non -AI tools that have rebranded as AI for sweet, sweet clout.
I’m reminded in this moment of AI that early mapping programs used to tell people to drive through fucking buildings as the most optimal route to a destination. They fixed that real quick, but folks clearly learned fuck all from the event.
“Are you smarter than an AI,” depends on what you mean. The AI does patterns faster than me but can’t tell me the meaning of those patterns for my life or their implications for others. The AI could predict the next possible action but it can’t tell me how to make the decision for myself.
If you know faculty in higher-ed this month, consider venmoing them some money, taking them to the local liquor store, or funding whatever hobby they use to preserve their sanity. In this age of AI slop and administrative imposition of AI we're going to need it.
Something that irks me is that if I convert a scanned text into ocr, which makes it searchable, this enables Adobe's AI slop to use its summarize feature on the text. This irks me because I make sure all my course texts are OCR'd for accessibility reasons.
I'm finishing up a manuscript on AI and the Promise of Happiness, and I need to find an editorial that past John foolishly did not bookmark. It was from the editor's desk of some pper and it basically said how they believed that journalists needed to learn to use AI? Any thoughts?
In Valen’s name, the fact that predictive text and autocorrect both try to correct “fucking” to “ducking” and choose to ignore fek”ihr, Grethor, and Khaless tells you everything you need to know about AI, even in its most simple form, as a disciplinary tool.
Center = Dr. Johnathan Flowers, Bisexual of the Blade (Alt-text or DIE). Left = members they follow (green edges). Right = members who follow them (blue edges). Top = mutual follows (orange edges, slightly larger). Drag any node to reposition; click to open that profile.