📣 Focus group participants needed: calling novelists, publishers, and literary agents! 🗣️ Have your say on how Generative AI could transform our industry by participating in this significant Cambridge University research, funded by BRAID UK and partnered with Institute for the Future of Work. 1/3

The reason it’s not getting more coverage is that remedying the conditions that led to this, which we’ve been warning about for years, would require unwinding dangerous x surveillant norms the current tech industry relies on. Coverage or no, the find out era has arrived and a

Most MIT profs who I know are extremely honest and morally upright. I hope this is an exception.

At #NeurIPS2024, the keynote speaker perpetuated explicit racist stereotypes against Chinese students. Generalizations against a community subject to discrimination, even as an “example”, further provokes discrimination. Below, video of an audience member’s perfect response.

It breaks my heart that scholars are sharing examples of their integrity and academic rigour in response to the racist assumptions against Chinese students and scholars that reared its head in Rosalind Picard’s keynote at #NeurIPS2024

I keep getting asked about my take on these CO2 estimates for the o3 model by the press and members of the community, so I'll interrupt my vacation to comment 🤓 TL;DR- any kind of estimate is a proxy, and instead of wasting our time and energy, we should demand accountability.

JOB OPPORTUNITY: Looking for a 2-year postdoc to join our project AIDEMOCRACY, examining Big Tech influence on AI strategy. Deadline is 28th Feb: jobs.gold.ac.uk/vacancy/postdo…

1/ One year ago, Sinica found a new home after The China Project ceased operations. It’s been a year of rebuilding, growing, and reaffirming my core mission. If you value thoughtful, in-depth discussions on China, I hope you’ll consider supporting my work. 🧵⬇️

📣New preprint!📣We’ve long known humans tend to anthropomorphize computers. But with the rise of social AI applications, like AI companions, studying this is now more crucial than ever. We introduce a new method for *empirically evaluating* anthropomorphic behaviors in LLMs🧵

UW’s Tech Policy Lab and I invite applications for a 2-year Postdoctoral Researcher position in "AI Alignment with Ethical Principles" focusing on language technologies, societal impact, and tech policy. Kindly share! apply.interfolio.com/162834 Priority review deadline: 3/28/2025

Free speech no longer exists in the US government. For The Atlantic I spoke with 12+ federal workers in 6 agencies who said the Trump admin’s actions have led to pervasive self-censorship, even on issues some view as critical to national security. 1/ theatlantic.com/technology/arc…

It's always interesting to me that despite how abysmally inaccurate these models are, their use is justified through alleged need for "speed". If you're willing to deploy a model with 25-50% accuracy for speed gains, just do away with the model and indiscriminately bomb at scale.

Looking forward to joining OxDEG - Oxford Digital Ethnography Group this Monday to share my ethnographic research on local data centre resistance in the UK. Sign-up below or join IRL in Oxford ☺️

Signal President Meredith Whittaker calls out agentic AI as having ‘profound’ security and privacy issues tcrn.ch/3DlyFVQ

There's a lot going on, but we should not let this slip under the radar—Google and OpenAI are cynically seizing on the Trump admin's hawkish, US-first AI policy instincts to make a play at gutting artists' copyright protections once and for all.

1/ I've been revisiting Jeffrey Ding's book *Technology and the Rise of Great Powers* ahead of a long-overdue podcast interview with him. His central argument is most persuasive: Tech diffusion and not leading-sector 0-to-1 innovation ultimately matters more in national power.

🚨 BREAKING Police have raided a Quaker meeting house and arrested six young people meeting to discuss concerns about the climate crisis and the Gaza genocide. On 27 March, more than 20 officers - some carrying tasers - reportedly forced their way into Westminster Meeting House

Datacenter capex is going to 0 Grok struggling deploying Anthropic extremely low rate limits for paid + Cursor Google not offering 2.5 on API OAI not serving o3, struggling to serve 4o image Azure, still not serving many models including o1, o3, deep research, 4.5

Excited to be speaking on AI safety evaluations Knight First Amendment Institute later this week with these wonderful speakers (and many more)! Come join us in person in NYC if you can or online :)

Whenever you see Palantir win a civilian government contract (I'm looking at you UK), consider the ways in which they'll use your private data towards these purposes.