Research Scientist (Frontier Planning) at @GoogleDeepMind. Research Affiliate @Cambridge_Uni @CSERCambridge & @LeverhulmeCFI. All views my own.
London
Joined May 2013
- Tweets20.8K
- Following2.3K
- Followers6.6K
- Likes178K
Pinned Tweet
Happy to announce I am now a Senior Research Scientist at Google DeepMind. I will be working with @AllanDafoe on his Frontier Safety and Governance Team, along with @JoslynBarnhart, Lucy Lim, and @_lewisho, as well as partner teams.
I’ll be working on ‘frontier planning’ – trying to figure out what new considerations could be coming with advanced AI and on the path to Artificial General Intelligence, advising Google DeepMind and Google, and collaborating with governments and society at large based on those insights.
We are at a moment of inflection - in the next five years, we have a rare window of opportunity to shape the governance of frontier AI. Only with sophisticated and nuanced governance can we push the boundaries of science, and discover new technical breakthroughs that benefit everyone.
Google DeepMind has a strong track record of proactive AI governance, investing in safety and security, and building AI tools that advance science and benefit humanity. It has exceptional and stable leadership: Nobel Prize Winner Sir @demishassabis makes me proud to be British; @ShaneLegg foresaw the promise and perils of scaling before almost anyone. I’ve worked with Allan’s team for years and respect and trust them deeply. I can’t think of a better team to work with on the ‘path to AGI’.
I’ve spent the past eight years at Cambridge University, at the Leverhulme Centre for the Future of Intelligence @LeverhulmeCFI and the Centre for the Study of Existential Risk @CSERCambridge. I am proud to have helped build those centres into a substantial, successful, permanent part of the University - both offering Masters programs to train the next generation of scholars. Together we have played a key role in building, establishing and making credible, mainstream and influential two new academic fields: existential risk and global catastrophic risk; and AI ethics, safety and governance. I will miss that wonderful intellectual home, but intend to remain closely involved as a Research Affiliate.
To find out more on our team’s work, you might be interested in recent interviews with Allan Dafoe, Demis Hassabis or Shane Legg; the recent update to the Frontier Safety Framework; and the paper Levels of AGI for Operationalizing Progress on the Path to AGI.
Haydn Belfield retweeted
Two greats join our bet:
- @dwarkesh_sp on the explosive-growth side
- @Thom_Wolf on my side
benjaminmoll.com/growth_bet/
Total pot now $700k 💰
Haydn Belfield retweeted
Huge. British Govt announces:
• £165m: National AI Biosecurity Initiative: Includes £100m for a new AI Biosecurity Lab, using AI to detect emerging biological threats.
• £120m: AI Security Institute (AISI): Additional funding to strengthen AI risk research and evaluation.
• £70m: National Exposure and Real-Time Vulnerability (NERV) Centre: Identifying vulnerabilities in critical infrastructure.
• £60m: National Cyber Shield: Strengthening Britain’s cyber defences.
• £60m: UK Sovereign AI: Expanding investment into physical and industrial AI.
• £60m: R&D Procurement Scheme: Supporting companies developing AI security and NHS innovations.
• £30m: Mayoral AI Partnerships: Helping local businesses adopt AI, create jobs and improve public services.
• £23m : AI Economics Institute: Research into AI’s effects on productivity, employment and living standards.
• £18m: UK Advanced Chip Design Academy: Developing domestic semiconductor design expertise and reducing overseas dependencies.
1/7 I have two simple tests for British AI, driven by our values. Is it safe? And is it ours to shape? 🇬🇧
Today, Chancellor @JohnHealey_MP is announcing a major, ambitious package of investment to strengthen Britain's defences against emerging threats, back British innovation, and build our technological independence.
Both Toms are so great, and verification is one of the most interesting and important topics in the world. Well worth a watch or listen!
I interviewed @tsmilton_amodo about the tech we'd need to verify a potential US-China deal on AI.
We discuss Amodo's prototype for inference verification, as well as the myriad ways you could use power monitoring, zero knowledge proofs and/or unilateral surveillance to understand how an adversary is using compute.
Though solving some kind of verification is central to many people's theories of victory, there are precious few people working on it currently. Tom is one such person!
Listen to the episode on Youtube, Spotify or Apple Podcasts.
You can also find more about Tom and Amodo's work at
ai-verification.com/
00:00:00 Using compute verification for an AI slowdown
00:06:08 Proving a datacentre is only running inference
00:13:22 Reading a GPU's workload from its power draw
00:23:35 What would it take to get verification deal-ready?
00:37:49 How realistic is verification?
00:46:07 Preparing for a rushed US-China deal
01:07:20 How good is AI at hardware design?
Incredibly interesting on agents, infrastructure and security
I came back to Google to work on AI infra. Instead we got pulled in to solving the problem of exploding use of agents across the company.
Announcing Agent Compute Engine: bringing planet scale agentic infra to Google services and now to Google Cloud.
cloud.google.com/blog/topics…
Listened to the first 5mins of this and they got the thesis of both books wrong oof
New Delphi Podcast with Michael Arrington (@arrington), founder of TechCrunch and Arrington Capital
Michael broke the news of Google buying YouTube, worked with Steve Jobs, and backed Uber and Airbnb early
We get into the Fourth Turning, why the AI race won't slow down, what separates great founders from everyone else, and why he still believes Bitcoin is the hard money the world needs
He also opens up about selling TechCrunch, burnout, reinventing himself through crypto, and learning to play to his strengths
00:00 Intro & the Fourth Turning
07:00 AI, China & the Race to AGI
17:00 Founder Psychology & Investment Red Flags
29:00 Building TechCrunch & Lessons from Steve Jobs
44:00 Selling TechCrunch & Backing Uber and Airbnb
53:00 Moving Into Crypto & Navigating Market Cycles
1:00:00 Bitcoin, AI & an Uncertain Future
1:09:00 Advice for Founders: Play to Your Strengths
Full episode
Haydn Belfield retweeted
I am happy to announce an opening for a student researcher on my team at GDM for 2027, based in London - for a project at the frontier of multi-agent coordination, in duration of 3 months with a possibility of an extension.
Haydn Belfield retweeted
American chattering classes love to hate our country’s financial services sector, but mark my words: the depth and versatility of U.S. capital markets will be our mightiest weapon in the singularity. Anti-finance sentiment, still fashionable on both left and right, is low IQ.
What a weird discussion.
16% of Americans are retired, and they're the wealthiest and happiest group in the USA.
It's clearly not some bizarre dystopia, it's normal and works well.
Former Anthropic researcher Jacob Coxon (@hilbertspaess) and Jon Stewart discuss AI, jobs disappearing, universal income, and whether tech CEOs are creating a dystopian future
See my colleague Stephanie's excellent thread and paper for more empirical data on this subject
xitter.cf/scychan_brains/status/…
Haydn Belfield retweeted
We’ve seen an incredible response to Gemini 4 Argon from users, developers & enterprises: nearly 25 trillion tokens processed in the first 72 hours after launch - 3x what Gemini 3 did.
Hear more about Argon and our commitment to the AI frontier from @sundarpichai ‘s conversation with @kantrowitz
New: I spoke with @sundarpichai about Google's commitment to the AI frontier, how close we are to RSI, balancing DeepMind's compute needs vs. Google Cloud profits, and much, much more.
Full Big Technology Podcast episode here -->
Chapters:
0:00 Introduction
1:14 Google’s Frontier AI Strategy
3:47 Do Frontier Models Really Matter?
6:35 The Future of AI Agents
7:39 Recursive Self-Improvement
12:58 World Models and the Path to AGI
14:56 Google’s Compute Dilemma and AI Infrastructure
19:36 Is AI Overbuilt?
23:35 The Future of the AI Interface
31:42 Google’s Answer to Muse and Dots
35:03 Apple, Space Data Centers, and AI Coding
41:42 Optimism and AI’s Future
V good slopcore
I came to realize people liked the french #slopcore without having the lyrics 😅 sorry!
So here is a version with english subtitles!
Haydn Belfield retweeted
Hey Argon, make me some awesome space timelapses, stitch them together in a video with title cards and a soundtrack
I've been doing some hobbyist astronomy with Argon. Here's one of my experiments, where I tasked Argon with reproducing the first black hole images (from 2017/18) using the same raw data.
Argon:
- pulled raw 230 GHz radio fringes (UVFITS) from the Event Horizon Telescope archives.
- built an interferometry solver using 3-dish closure phases to cancel atmospheric noise
- started each run from a plain Gaussian blob with no ring templates or masks
- recovered M87* (2017 & 2018) and Sgr A* at a 91% to 98% match of the official EHT images
Haydn Belfield retweeted
We’re announcing the 17 selected organizations for the Google.org Impact Challenge: AI for Science—receiving $36M in funding + Google.org Accelerator support to tackle critical challenges in climate and health.
Congratulations to all 17 recipient teams! I am excited to see how these interdisciplinary efforts push the boundaries of scientific discovery for societal impact.
Meet the 17 selected organizations and their work: goo.gle/gic-ai-science
Meet the 17 recipients of the Google.org Impact Challenge. These teams are taking on some of science's hardest problems, from disease prevention to mapping the natural world.
Learn how they're unlocking scientific breakthroughs: goo.gle/gic-science
Got to close those loops
We are bringing agents to the wet-lab bench at Google Research to navigate complex biological landscapes. Learn how we are grounding agents in physical lab robotics to run fast closed-loop molecular design cycles, ultimately discovering single-digit nanomolar protein binders for cancer targets →goo.gle/4B5mMMB
Just ran this and it said "the innate genius of the British people"
💯
just read the model cards guys
AI ppl loooove to pretend they hold tons secret knowledge. Everyone has Seen Things, and have been trusted with the Esoteric Knowledge. It's also cringe if you don't partake because it implies you're not in the Inner Circle. In reality most people are kinda confused, working it all out ad hoc, relying on the same rumours, and so on - but the status games are hard to resist!
