Timothy Nguyen
@IAmTimNguyen
Machine learning researcher at @GoogleDeepMind & mathematician. Host of The Cartesian Cafe podcast. All opinions are my own.
Quantum field theory textbooks have been lying to you or have left you confused. For a typical passage like the following from Peskin&Schroeder - what does it mean to do a change of variables on an ill-defined path integral? For perturbative QFT, my paper resolves this issue: 🧵
Quick thread on the recent IMO results and the relationship between symbol manipulation, reasoning, and intelligence in machines and humans:
2025: The year where a gold medal at the IMO is won by an AI system. What a time it is to be alive and to be a mathematician! Huge kudos to friends and colleagues.
Very excited to share that an advanced version of Gemini Deep Think is the first to have achieved gold-medal level in the International Mathematical Olympiad! 🏆, solving five out of six problems perfectly, as verified by the IMO organizers! It’s been a wild run to lead this…
What happened after Sputnik? @ATabarrok provides a striking summary: NSF funding tripled in a year and increased by a factor of more than ten by the end of the decade. The National Defense Education Act overhauled universities and created new student loan programs for foreign…
Today, we at @OpenAI achieved a milestone that many considered years away: gold medal-level performance on the 2025 IMO with a general reasoning LLM—under the same time limits as humans, without tools. As remarkable as that sounds, it’s even more significant than the headline 🧵
1/N I’m excited to share that our latest @OpenAI experimental reasoning LLM has achieved a longstanding grand challenge in AI: gold medal-level performance on the world’s most prestigious math competition—the International Math Olympiad (IMO).
Speaking as a past IMO contestant, this is impressive but misleading - gold vs silver is meaningless, 1 pt below gold vs borderline gold is noise The impressive bit is using a general reasoning model, not a specialised system, and no verified reward. Peak AI maths is unchanged
1/N I’m excited to share that our latest @OpenAI experimental reasoning LLM has achieved a longstanding grand challenge in AI: gold medal-level performance on the world’s most prestigious math competition—the International Math Olympiad (IMO).
The problem with Eric Weinstein is that his theory has already been discredited, per my article hosted on Sabine's blog. Odd that you left that part out @skdh but instead praised his supporters as if it were some noble heroism. backreaction.blogspot.com/2021/03/guest-…
Had a lot of fun being part of the inaugural @MLStreetTalk AI Meetup, where I got to talk about my work on Understanding Transformers via N-Gram Statistics, discuss LLM benchmarking with other panelists, and get testy with philosopher Mark Bishop on theory of mind.
Really enjoyed discussing the state of AI benchmarking alongside Prof Mark Bishop, @IAmTimNguyen, Enzo Blindow & @ecsquendor at @MLStreetTalk's first in-person event in London yesterday. Looking forward to many more!
Understanding Transformers via N-gram Statistics 🔍 On July 9th, @IAmTimNguyen dives into how transformer predictions often align with simple stats—unpacking learning dynamics, overfitting detection & what it means for model generalization. Learn more: tinyurl.com/9ts5bj7y
Join @MLStreetTalk for its first AI/ML meet-up in London! The night will include talks and panels featuring @GoogleDeepMind's @IAmTimNguyen, @Cohere’s @max_nlp, MLST host @ecsquendor, and Prolific's Enzo Blindow. Connect with AI/ML pros and grab some refreshments on us. RSVP 👇
Looking forward to attending and giving a talk at @MLStreetTalk's first social here in London. Come join to grab a drink and mingle with other AI researchers!
We are running our first physical event in London on 14th July! We have Tim Nguyen @IAmTimNguyen from DeepMind and Max Bartolo @max_nlp from Cohere and Enzo Blindow (VP of Data, Research & Analytics) at @Prolific joining us. Not many seats for the first one.…
✨ New blog post 👀: We have some very fast AI-generated kernels generated with a simple test-time only search. They are performing close to or in some cases even beating the standard expert-optimized production kernels shipped in PyTorch. (1/6) [🔗 link in final post]
Today marks a big milestone for me. I'm launching @LawZero_, a nonprofit focusing on a new safe-by-design approach to AI that could both accelerate scientific discovery and provide a safeguard against the dangers of agentic AI.
Every frontier AI system should be grounded in a core commitment: to protect human joy and endeavour. Today, we launch @LawZero_, a nonprofit dedicated to advancing safe-by-design AI. lawzero.org
Looking forward to the East Anglia 2025 Festival on June 14-15 where many thinkers and artists will come together and discuss some of the most exciting ideas of today. I'll be on a panel with mathematician @MarcusduSautoy and human rights layer Susie Alegre on whether the AI…
This has been debunked: x.com/IAmTimNguyen/s… Looking out for you @seanmcarroll. CC: @piersmorgan
What really happened on @PiersUncensored? You’d *never* believe it if I told you. Here: