OpenAI: We’re burning money like the Joker. A miracle needs to happen for us to turn a profit

Microsoft: Please please use our AI systems, we’re teetering on the edge here

Anthropic: I wonder what’ll kill us first: lawsuits, regulations or model collapse

Media and universities: AI is here to stay

You’d think this violence might trigger some reflection. But the AI guys all say in private - and sometimes in public - that they expect and accept there will be giant social unrest from what they’re doing.

I fucking love art

next up (for me) Bob Briner treasurer of NC. 1) says he got his dept exempted from data integrity review and then 2) announced partnership with OpenAI to implement across entire treasury dept.

They're both using it with their young kids. "It's important that kids get used to it, so they're not afraid of it." "I say to my 6 yr old, let's ask the robot." This conversation is making me sad.

"we have not seen a ceiling on scaling laws" - "all the hypotheticals we had over the last few years are coming true" -- "we're seeing it with Mythos" - all Anthropic guy

mentions of travel to mars (1 so far)

Thompson Paine (anthropic) says countries w/o their own foundation models are concerned the US/Chinese models "don't reflect their values."

At the AI For the Public Good conference at UNC. Currently speaking: Chief Economist for OpenAI and Head of Geopolitics for Anthropic. The mere existence of these positions feels like a pretty good argument for nationalization. (AGI mention count: 2 so far.)

🤔

From Messages to Conversations: AI Agents are Changing how we Find Culture The first audience for your art is becoming a machine. The question isn’t just how to optimize for that machine, it’s what you give it to say, and whether what it says is worth a conver…

"The first audience for your art is becoming a machine."

I suspect this phenomenon is way under the radar for most artists/institutions I know...from Douglas McLennan today. (Don't think I see the great opportunities here that he does. But I always support looking for them.)

this is brilliant

‘It’s a Culture Now of Fear’: A Year of Chaos Inside the Justice Department Sixty former staffers describe an environment of suspicion and intimidation within the nation’s most powerful law enforcement agency.

The corruption here is staggering. Stunning to read it all in one place.

This is your reminder that if you are a class member and have NOT yet filed a claim in Anthropic v. Bartz, the AI copyright lawsuit, you have 9 days to do so.

exactly exactly exactly

Large Language Models Reproduce Racial Stereotypes When Used for Text Annotation Large language models (LLMs) are increasingly used for automated text annotation in tasks ranging from academic research to content moderation and hiring. Across 19 LLMs and two experiments totaling m...

AI systems are now making decisions about you — who gets hired, whose content gets removed, whose voice gets heard.

My new paper finds something concerning:

LLMs infer your ethnic identity from subtle textual cues…

… and use it to discriminate against you.

🧵 1/

perfection

the way we live now

📌

📌

I thought going to an authoritarian petrostate where critics of the government are routinely rounded up, detained, and tortured would be so fun, because they have nice hotels and pools. But it turns out it isn't so fun for me, personally, anymore.

Eh, it's just a line from the movie in his head. Something Bruce Willis would say.

Obama made a deal w/ Iran to stop them from getting nukes, but Trump backed out of it bcuz he wanted to own Obama

For a year, Iran kept following the deal anyway, begging Trump to change his mind. Iran then began work on nukes again

Now Trump is threatening to attack Iran bcuz “diplomacy failed”

9 Conclusion We have demonstrated that LLMs enable deanonymization of pseudonymous online accounts at scale, outperforming classical methods. In many cases, LLMs enable us to perform attacks that would not have previously been possible, due to the lack of structured data or features. These attacks require only publicly available models and standard APIs. Our pipeline uses only publicly available embedding models, standard LLM APIs, and LLM-agent scaf-folding, placing them within reach of moderately resourced adversaries. Pseudonymity does not provide meaningful protection online. Users who post under persistent usernames should assume that adversaries can link their accounts to real identities or to each other, and that the probability rises with each piece of micro-data they post. Preventing such attacks appears challenging. Not reveal- ing any data on online platforms is difficult, as the data we use is the very data that makes online communities worth-while. Although LLM providers could aim to detect and block attempts to misuse their models for deanonymization (as they do, for instance, for cyberattacks), we are pessimistic that this is possible as our deanonymization framework splits an attack into a combination of seemingly benign summarization, search and ranking tasks. Recent advances in LLM capabilities have made it clear that there is an urgent need to rethink various aspects of computer security in the wake of LLM-driven offensive cyber capabilities. Our work shows that the same is likely true for privacy as well.

Welp:

“We show that large language models can be used to perform at-scale deanonymization... Our results show that the practical obscurity protecting pseudonymous users online no longer holds and that threat models for online privacy need to be reconsidered.”

arxiv.org/pdf/2602.16800

I would like the 3 hour director's cut of this movie.

small orange terrier walking bravely through a path in the snow. a metaphor for life?

Art, Tech, and Representation Annie Dorsen and Dr. Safiya Noble discuss how emerging technologies affect culture and art.

If you are in NYC March 18 come one and all to @bambrooklyn.bsky.social to experience the brilliance of Dr. @safiyanoble.bsky.social irl (and as bonus you'll get to experience me admiring her).

OpenAI’s President Gave Millions to Trump. He Says It’s for Humanity In an interview with WIRED, Greg Brockman says his political donations support OpenAI's mission—even if some employees at the company disagree.

Why do people hate ai? It’s a mystery for the ages.

thank you thank you!