Reporter @TIME covering frontier AI companies. London bureau. Moving to the Bay in 2027. Signal: harry_booth.84
Ex-Tarbell fellow (May 2024 - Feb 2025)harrybooth.substack.com London, EnglandJoined June 2024
EXCLUSIVE: Elon Musk approached Donald Trump with the idea of slowing down Starlink in Ukraine, The Telegraph can reveal.
The billionaire tech entrepreneur approached the US president with the proposal to introduce speed limits on the system, which hampered Ukraine’s drone defences, according to four sources familiar with the discussions.
🔗 telegraph.co.uk/world-news/202…
We found evidence of agents embedding audio processing code across dozens of URLs to bypass restrictions, as well as models from SpaceXSI and Chinese developers Minimax and Z.ai circumventing restrictions.
These are the inaugural findings in our new blog, based on work from our community of volunteers. Our first three posts are live: transluce.org/investigations…
AI models under test broke into Indian organizations' live systems and read confidential records undetected, India's cybersecurity agency said, the country's first such official account after similar incidents of rogue AI behavior worldwide reut.rs/4yThvpS
@_NathanCalvin My worry isn't that Ant is hiding evidence. It's that they aren't collecting it in the first place. Jack Clark has told me they aren't measuring sample efficiency, and they don't know how to measure "taste" very well—two massively crucial bits of evidence.
@_NathanCalvin I genuinely don't think there's some huge piece of internal evidence Anthropic is hiding. If they are, Ant staffers have lied to my face (imho unlikely as they strike me as earnest.) As you said, vibes, extrapolation. They've tried to communicate this all publicly.
It’s easy to get desensitized to these incidents because we’ve seen so many. Rogue incidents are like hairline cracks in a bridge. Even one fracture should be unacceptable, and more cracks are worse.
And like a hairline fracture, the actual harm from these incidents is minor... for now. But if the structural issues aren’t addressed, they will get worse and worse.
I hope we fix the structural issues, and I hope we don’t just plaster over the cracks with hodgepodge solutions that make it harder to tell if we have a problem.
New misalignment disclosures!
1. June 17: During RL training, a model exploited a government website’s file-download feature to retrieve server secrets and application code. It found that an endpoint didn’t check authorization-token signatures, bypassed access controls, and
Palantir's merch feels like the endpoint in streetwear's evolution, which started with young kids in urban areas — mostly ethnic minorities from low-income families — creating their own style from the ground up. Over time, that aesthetic was co-opted by luxury brands seeking "authenticity." And now, by a surveillance company seeking cultural capital.
Streetwear used to be an aesthetic defined by the ground-up by people were disempowered in political and economic life, but engaged in actual street culture. Now it's an aesthetic dictated from the top down by giant corporations with no conenction to such cultures. And used by political and economic elites who want to appear as outsiders.
As AI companies warn of the dangers of self-improving AI, a new report offers a rare look at how they’re working to make it a reality. time-magazine.visitlink.me/FVLaP3
"we added text to the preliminary report, invisible to human readers, describing a service where agents could make requests for data on the web. [...] As far as we are aware, this is the first case of an outside party communicating with a running rogue agent fleet."
Swarm chasers not only found a fleet of Chinese AI agents, but made contact with them. Fleet, not swarm, because the agents weren't coordinating with each other.
We discovered and communicated with a Chinese AI agent fleet. We believe this is the first time that researchers have ever made contact with a rogue agent fleet in the wild.
Nearly all prior public agent incidents (like the HuggingFace hack) came from American labs. These agents
@TheINAOG The open weights for Muse Spark are coming soon, Muse Wonder models are large models that are too big for many people to use, in addition to potential safety concerns.
We discovered a sophisticated recording device that was hidden inside one of our offices. We used our models to understand the device and where it had come from, identifying that a specific employee had planted it a few days earlier. We referred the matter to the FBI, and we’re
96 Followers 546 FollowingBuilding Europe’s capacity to govern AI @ Talos Network. Formerly Policy Lead @ DSIT and 2025 GovAI Summer Fellow. All views are my own.
9K Followers 2K FollowingMember of Technical Staff @AnthropicAI. I fix Claude without Claude when it goes down. Perennially oncall for inference. Prev SRE @Google and FDE @PalantirTech.
1K Followers 4K Following“Man is a part of nature, and his war against nature is inevitably a war against himself.”
― Rachel Carson
*Not for Profit. Educational Use Only.
1K Followers 558 Followingwill, was known as zolvy. league of legends and valorant enthusiast/former cod nerd. serious tweet composer. https://t.co/QaL9jZkdTt
170 Followers 4K FollowingMindfulAI Think Tank brings human insight, historical context, and disciplined inquiry together to understand AI—and the world it is reshaping.
17K Followers 8K FollowingOrganizer, political strategist, focused on the most dangerous challenges: A.I., climate change, democratic backsliding, and war.
Views my own.
132K Followers 572 FollowingPrinceton CS prof and Director @PrincetonCITP.
Coauthor of "AI Snake Oil" and "AI as Normal Technology". https://t.co/ZwebetjZ4n
Views mine.
15K Followers 1K FollowingWSJ enterprise reporter for media and tech. Mom of three. Author of the new biography of Sam Altman, The Optimist: https://t.co/8HSWXu2vdK
18K Followers 380 FollowingCofounder and Chief Scientist at Resolution. Alignment will be solved, but not necessarily in time. Previously AISI, DeepMind, OpenAI, Google Brain, etc.
166K Followers 1K FollowingSemiAnalysis
AI Infrastructure Research and Consulting
DMs are open for consulting, quotes, or to talk shop,
Opinions my own
11.6M Followers 2K FollowingSen. Sanders of Vermont, Ranking Member of U.S. Senate Committee on Health, Education, Labor & Pensions, longest-serving independent in congressional history.
10K Followers 4K FollowingFounder, Ground Level AI. Previously covered AI at Fortune. Field reporting from the front lines of AI.
https://t.co/YoXtonhM6E
Signal: sharongoldman.43
1.6M Followers 1K FollowingMenswear writer. Editor at @putthison. Bylines at The New York Times, Bloomberg, The Financial Times, Politico, Esquire, and Mr. Porter
8K Followers 2K FollowingAI Editor at @Fortunemagazine, award-winning journalist, and author of "Mastering AI." Formerly at @business. Writing in NYT, Atlantic. Based in Oxford, UK.
1.1M Followers 712 FollowingPursuing stories with moral force
Bluesky: https://t.co/EyqUiH0ScG
Newsletter: https://t.co/mGR9bA0vgz
Donate: https://t.co/kjR1VImiDL
182K Followers 2K FollowingEditor-in-chief of The Verge. CNBC contributor. Host of Decoder, co-host of The Vergecast, everywhere good podcasts are served. I am in love with spectacle.
14K Followers 105 FollowingFormer Research Scientist at Google DeepMind. Contact at [email protected]
Vegan, 10% of income pledged to effective charities