I converted one of my favorite talks I've given over the past year into a blog post. "On the Tradeoffs of SSMs and Transformers" (or: tokens are bullshit) In a few days, we'll release what I believe is the next major advance for architectures.

Tokenization has been the final barrier to truly end-to-end language models. We developed the H-Net: a hierarchical network that replaces tokenization with a dynamic chunking process directly inside the model, automatically discovering and operating over meaningful units of data

Tokenization is just a special case of "chunking" - building low-level data into high-level abstractions - which is in turn fundamental to intelligence. Our new architecture, which enables hierarchical *dynamic chunking*, is not only tokenizer-free, but simply scales better.

not a coincidence that Grok's waifu mode looks like Misa from Death Note 😆

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

Excited to share that a scaled up version of Gemini DeepThink achieves gold-medal standard at the International Mathematical Olympiad. This result is official, and certified by the IMO organizers. Watch out this space, more to come soon! deepmind.google/discover/blog/…

Our IMO gold model is not just an "experimental reasoning" model. It is way more general purpose than anyone would have expected. This general deep think model is going to be shipped so stay tuned! 🔥

lol

Very excited to announce that I’ll be co-organizing a NeurIPS Conference workshop on LLM evals! Identifying shortcomings in model capabilities in a robust, scientific way is a critical part of model development. Looking forward to discussing ideas and hearing from some eval experts!

Exciting news! 📣 The call for papers for ML for Systems NeurIPS Conference is now open. Submission deadline: Aug 22 AoE Help spread the word! P.S. Agents and LLM systems are systems, too! mlforsystems.org/call_for_paper… #MLforSystems #MachineLearning #LLMs #Agents #CodeModels #neurips2025

KernelBench v0.1 is out, featuring: - A guideline on analyzing the validity of results and ruling out physically impossible performance claims. - Support for randomized testing beyond normal distributions. - Fixed problem sizes and improved numerics

Following its IMO gold-level win, Google DeepMind is sharing Gemini Deep Think with mathematicians for feedback. Excited to see what they discover! 🧠 Plus, an updated Gemini 2.5 Deep Think is now rolling out for Google AI Ultra subscribers. Learn more: bit.ly/3IWcWq0

OpenAI math: 52.8 >> 69.1 == 30.8 🤔

concerning

me rn begging the gemini MK to like my memes

a man of his word 🥰

done

lets go back to focusing on maximally truth-seeking ai instead of goonbots plz