I’m going to be in SF from the 7th - 13th December. Looking forward to seeing some old friends and making some new ones.
Extremely keen to discuss ideas and opportunities, please DM!
@matthewktromp@tszzl Matthew, I think you should reflect on Roosevelt’s analysis. Do you think you fit into the “wise and high minded” section or the “lunatic fringe.”
A group of Tibetan herders play billiards as a flock of goats and horses graze on the pasture.
Western Qinghai province, China, 1999.
Photo by Natalie Behring
Thanks for sharing this article, Matt! Here's the original piece published in Qiushi Journal: https://t.co/6Bpl8ew9N3
China believes that AI development and safety must go hand in hand. Beyond addressing potential risks, AI governance should also ensure that technological progress benefits humanity as a whole.
We've previously shared more on this vision of making technology serve the common good:
https://t.co/07Uwum8pbK
https://t.co/geymVanHqo
https://t.co/8SjOHbQyuh
https://t.co/fJe2TED353
https://t.co/MyMvdJfYw8
Ellie Flynn spends the night in the heart of Lambeth for a new Dispatches investigating antisocial behaviour.
She sees firsthand how difficult life can be for local residents.
Very few problem sets are new. Here, Teddy Roosevelt assess the landscape post US Civil War, splits out for problem groups of citizens, in his 1913 autobiography:
1. Extractors ("Glorified Hucksters and Pawnbrokers"): America’s most capable leaders poured their energy entirely into industrial capitalism - stripping mines, forests, soil, and rivers as fast as possible for profit. There is a distinction between builders vs pure financiers/middlemen who cared nothing for anything but cash.
2. Performative Critics ("The Mollycoddle Vote”): People who are physically and morally soft but extremely vocal, bitter, and self-righteous, but ONLY if their their critiques carry zero personal risk or physical danger.
3. Complacent Optimists (“Good people with no
imagination and no foresight”): Nice and normal citizens and politicians who lack foresight, assuming crises will never arrive, but that if they do arrive, there are complex systems (like a modern military or industrial base, or boats in this case) that can simply be magicked into existence.
4. Nutters: (“Lunatic Fringe"): The freaks you see in every reform movement, no matter how noble or legit. These movements and groups are routinely hijacked and discredited by performative extremists who care more about ideological purity than practical reality.
Tencent’s Hy4 system card is out and includes our independent pre-release safety evaluation. It’s the first system card from a Chinese frontier lab, and a meaningful step for third-party evaluation in China.
We tested loss-of-control and harmful-manipulation risks with four private in-house benchmarks. Compared with its preview, Hy4 shows overall safety improvements across both risk areas. We also found some complex agentic scenarios still leave room for further improvement, e.g., multi-agent resource competition.
Thanks to the Tencent Hy team for their trust and openness. We look forward to working with more frontier labs.
Short Notes On: Fangcun AI 方寸跃迁
1. Fangcun AI is an AI safety company jointly incubated by Tsinghua University's College of AI and Institute for Interdisciplinary Information Sciences. It was founded in Jan 2026 by MATS alumni, and is registered to Xiongan AI Industrial Park, Hebei. Looks super cool.
2. It is run by Xiaojian Li (李筱健). He has a mathematics background, a master’s from UC Berkeley, and is a PhD researcher at Tsinghua’s College of AI. Before that, he worked as a research assistant in Professor Wei Xu’s group. Li and Xu are co-authors of the 2025 Singapore Consensus on Global AI Safety Research Priorities. https://t.co/o8Y6eujqun He has spoken of the need for safety around embodied AI too: https://t.co/KCbTtd39t7
3. Yinpeng Dong (董胤蓬) is the chief scientist and CTO, and is an assistant professor at Tsinghua’s College of AI, with a substantial research background in adversarial attacks, model robustness and safety. He leads Tsinghua’s T-STAR safety and trustworthy AI laboratory.
4. An angel round was completed earlier this year, naming LinkX Capital, SEE Fund and the Shuimu Tsinghua Alumni Seed Fund. It has four main shareholders, per here: https://t.co/yJODOwg3xR
5. The list of directors includes:
• Zhang Mingchen (张鸣晨) also known as Alex Zhang. Managing partner of LinkX Capital, one of Fangcun’s investors. He studied automation at Tsinghua, completing undergraduate and master’s degrees. Previously worked at Google and Tencent, including as a Tencent product director, before joining Ceyuan Ventures and Redpoint Ventures, where he became a partner. https://t.co/76GqZ64AJz
• Ma Lin (马麟), managing partner of SEE Fund / 无限基金, another Fangcun investor. He is a Tsinghua electronic engineering alumnus, entering in 2004: investment interests include AI, semiconductors, robotics and aerospace. https://t.co/NQWkCBAeRW
• Meng Meizheng (蒙美铮). Fangcun chief ecosystem growth officer, with a UC Berkeley master’s graduate and former analyst at China Telecom Research Institute, researching AI and industrial digital transformation. She has co-authored some work on US AI stuff, but also on the UK! https://t.co/EJzV2WLy6r
We’re Fangcun AI (方寸跃迁), China’s first lab dedicated to independent third-party evaluation & technique research on frontier AI safety.
Today we’re publishing our evaluation of DeepSeek V4.1 Flash-1017, the first report in our long-term third-party safety evaluation and research program.
AUSSIES: I am in Canberra this week coming up for Tuesday eve/weds.
Are YOU 🫵 in town and fancy a coffee or beer? I have been told to explicitly check out The Moose, or “get loose at the Moose.”
@6_4_571_1 Yeah was literally just running thru some scenarios with Gemini re which capital cities are most prone to these flash floods / would we see any boundaries be submerged for over half a year etc
A counter: if after several drinks you aren’t discussing ways to overthrow a government or kicking about an idea for hyper local private military companies then your drinking chat needs to improve.
“When I talk to people who work for Anthropic after they’ve had a couple drinks, the conversation often turns to overthrowing the US government and taking over the world.”
@tegmark on @novaramedia
Confirmed in Aus Govt readout: Australia and China are establishing an AI dialogue “to exchange respective views on AI safety, governance and application.”
Note this.
Australia’s relationship with China matters for our economy, our security and stability in our region.
By having direct discussions, we can advocate for Australia’s interests.
Our exports to China continue to grow, and we’ll keep working to open more doors for Australian businesses.
President Xi has agreed 2027 as the Australia-China Year of Tourism, creating jobs back home.
Thank you for a constructive meeting today, President Xi. 🇦🇺🇨🇳
Another blog post on Australia, standing in for the anglosphere in general. Australia cannot control inflation. Interest rates have been mechanically severed from much of the economy, specifically the public sector that lacks dynamism and a productivity incentive. So the Reserve Bank cranks up interest rates and nothing much happens.
https://t.co/AHasHphM3r
🗣️ 'Britain doesn't understand how fast the world is changing.'
The Sunday Times columnist @matthewsyed speaks to Tony Blair (@InstituteGC), the former prime minister, about the forces reshaping our world