dan browne

Bruce Schneier on the Agentic AI trilemma: “Fast, smart, secure; pick any two. Fast and smart—you can’t verify your inputs. Smart and secure—you check everything, slowly, because AI itself can’t be used for this. Secure and fast—you’re stuck with models with intentionally limited capabilities.”

Adversarial Poetry as a Universal Single-Turn Jailbreak Mechanism in Large Language Models We present evidence that adversarial poetry functions as a universal single-turn jailbreak technique for large language models (LLMs). Across 25 frontier proprietary and open-weight models, curated po...

Next frontier in AI safety is.....adversarial poetry?

arxiv.org/abs/2511.153...

Started a thread in the other place and bringing it over here - I really think we should be more vocal about the opportunities that lay at the intersection of these two options!

So I'm starting a live thread of new roles as I become aware of them - feel free to add / extend / share :

if you were to teach a class on the pre/history of AI in terms of key concepts ideas, what would they be? the mind/body problem? abstraction vs materialism? history of the database? automation?

People need to start investing in, like, Polaroids and Technicolor film if you ever want to use photographic or video evidence in a courtroom after 2026

a colorful visual alphabet labeled AI Ethica Alphabet, with letters, icons, and brief explanations. Examples include: B is for Bias with "Consent: AI can reflect and amplify unfair prejudices"; G is for Governancs with "Governance: Rules and oversight for safe i deployment"; G is for Governance with "Justice: injured indegeend If train ing snhzenic data repeordity"; P is for Regulation with "Oversight: Humans must se-powy important AI tasks"; R is for Regulation with "Guality Data: Good data prevents poor or biased outputs"

There's a trend going around TikTok to have ChatGPT create visual alphabets (because the output is always full of ridiculous mistakes). I generated this one to use an example to help explain WHY AI image generators make these kinds of mistakes. It turned out even better/worse than expected.

This neologism rubs me the wrong way. Memes aren’t geologic epochs.

I am so damn excited to share my new book, Law and Technology: A Methodical Approach. Full of stories and priced at $40, this book explains why law finds technology so difficult to regulate, and what to do about it. global.oup.com/academic/pro...

A bumbling game of robot soccer was a breakthrough for embodied AI The soccer tournament was played by AI-powered robots in Beijing on Saturday. China is investing tens of billions of dollars to extend its lead in next-gen robotics.

"China’s robotics market is projected to grow at an annual rate of 23% and reach $108 billion by 2028... By 2050, China is expected to have 302.3 million humanoid robots in use, far ahead of the U.S. projection of 77.7 million" www.nbcnews.com/world/china/...

Read this thread ⤵️

Computer-vision research powers surveillance technology - Nature An analysis of research papers and citing patents indicates the extensive ties between computer-vision research and surveillance.

New paper hot off the press www.nature.com/articles/s41...

We analysed over 40,000 computer vision papers from CVPR (the longest standing CV conf) & associated patents tracing pathways from research to application. We found that 90% of papers & 86% of downstream patents power surveillance

1/

Stay cool out there

Building AI Felt Like Watching An Alien Intelligence Arrive: Early OpenAI Researcher OpenAI had stunned the world by releasing ChatGPT in November 2022, but it turns out that the researchers working on the technology were even more stunned at what they'd developed. Jeff Clune, a fo...

How it felt to know AGI is coming soon long before the world was paying attention.

officechai.com/ai/building-...

This story is too dumb not to be a plant

Our crisis is not loneliness but human beings becoming invisible | Aeon Essays Our crisis of work and technology is one in which too many people feel that nobody sees them as a fellow human being

“One major culprit behind the wave of depersonalisation is the widespread reduction of individuals to data. Feeling invisible can stem from repeatedly experiencing standardised interactions”
aeon.co/essays/our-c...

www.nytimes.com/2025/05/30/o...

Google search "do apples contain lycopene?" The AI overview says "No, apples do not contain lycopene" whereas the highlighted search result begins "Apples contain lycopene"

the modern information environment

Douglas Adams quote about the Electric Monk

Douglas Adams saw it all coming.

DuoLingo: we're not going to cut any jobs, but we want to use AI to limit how much more hiring we have to do while still growing and addressing "bottlenecks" that used to make us hire consultants.

AI panic monkeys: DuoLingo are cutting jobs!!!

#AIEthics #FutureOfWork #FoW

Bar graph titled "Top 20 Countries by Research and Development Spending." The United States leads significantly in spending, followed by China, Japan, and Germany. Other countries include the United Kingdom, South Korea, France, and Australia, among others.

To see why destroying the US science enterprise is so detrimental to the world, we have to realize that the US spends more on research and development than any other country in the world (by far), and if we remove China and Japan, the US spends almost as much as ALL other countries combined.

rarely has there been a purer example of human progress created by someone repeatedly pressing a button labeled “DO NOT TOUCH”

I get at least three heartbreaking, incoherent emails a day now from random strangers who believe they’ve awakened an LLM god or LLM consciousness of a new dimension. The most salient quote from the screenshots below: “(An LLM) will never just say, ‘Jesse, what the fuck are you talking about??’”

It took me a very long time to learn "do your best" does not mean "give 100% effort even at the expense of your physical or mental health" or "give A+ effort even when it is clearly not appreciated or valued". Sometimes doing your best is heating up a can of soup and going to bed early.

This is a phenomenal essay to re-read closely today.

Thanks for all your hard work @jagmeetsingh.ca

Important read. Everything must be understood through the political economy of collapse.

Ghost Work by Mary L. Gray and Siddharth Suri

Boycotting the grocery store until they start buying an equivalent volume of groceries from me

pro tip: if you can afford to blow $250 million on a losing campaign in Wisconsin, you can afford to pay a 90% top tax margin