Pushmeet Kohli

One of @GoogleDeepMind's key aims is to unlock scientific progress on problems important for society. Generating clean energy is one of the biggest challenges of our time, and technologies like nuclear fusion could be the answer. While we continue to conduct our own research on this problem, our team has also been talking to experts across the field to understand the main AI opportunities and data obstacles, and to share recommendations to unlock faster progress. This kind of ‘AI Data Stocktake’ exercise can serve as a blueprint that policymakers, science funders, and industry can apply to many other scientific domains. Read more at: deepmind.google/public-policy/…

I am happy to introduce AI co-clinician, @GoogleDeepMind's research initiative to explore how AI could better amplify doctor's expertise and help deliver higher quality care to patients. We’re excited about our early results, and are taking a phased approach to our research explorations with academic and research collaborators. Read more in our blog: deepmind.google/blog/ai-co-cli…

Incredibly rewarding to see foundational AI research translate into tools that millions of people rely on. @GoogleDeepMind's AlphaEarth model is now powering high-resolution precise pollen maps for @googlemaps and Pixel Weather in the US - just in time for allergy season! 🌳🤧 Here’s an example showing pollen-producing Maple trees in New Jersey

Just back from an inspiring trip to Singapore 🇸🇬 where I met Theodore Kwan, an 8-year-old! chemistry whiz who is already using AlphaFold to explore and learn more about conditions like cystic fibrosis with his mum. It’s great to see @GoogleDeepMind's AlphaFold model inspiring the next generation of scientists :) Read more about him at: straitstimes.com/singapore/pare…

Thanks for the feedback @jamesproud. Really happy to see AlphaEvolve enabling you and your team at @substrate to discover new and more effective solutions. This is exactly why our teams @GoogleDeepMind and @googlecloud are so invested in the future of AI agents.

Using AI to solve the Traveling Salesman Problem at warehouse scale. 📦 AlphaEvolve helped FM Logistic improve its routing algorithm by 10.4%, resulting in a reduction of total warehouse travel by over 15,000 km per year. 🚚 A great example of how @GoogleDeepMind and @googlecloud are using AlphaEvolve to help companies become more efficient. Read more at: cloud.google.com/blog/products/…

Gemma 4 outperforms models over 10x their size! (note the x-axis is log scale!)

The true promise of AI lies in its ability to serve people. It was an honor to be invited by High Commissioner @VDoraiswami to speak alongside @RishiSunak and other esteemed speakers to share my reflections from the AI Impact Summit in India. 🇬🇧🇮🇳 As two of the world's most historic democracies, the UK and India have a unique opportunity (and frankly, a responsibility) to lead the way in ensuring AI works for the benefit of all humanity. To achieve this, we must prioritize AI's application in advancing Science, Education, and Healthcare —a mission central to our work at @GoogleDeepMind. By focusing on these pillars, we can move toward a world defined by health, security, and meaning.

Read our Nature paper: nature.com/articles/s4158… Read our IMO 2024 blog: deepmind.google/blog/ai-solves…

Our AlphaProof paper is in this week’s issue of @Nature! In 2024, @GoogleDeepMind's proof agents AlphaProof & AlphaGeometry together made a substantial leap in AI by achieving the silver-medal standard in solving IMO problems. The Nature paper describes the technical innovations required—in particular, the RL loop bridging natural language & symbolic rigor—that made AlphaProof possible.

At @GoogleDeepMind, we believe AI is the ultimate catalyst for science. 🧬 The best example of this has been the AlphaFold database (AFDB) of protein structure predictions which has been used free of cost by more than 3.3 millions researchers across the world! Today, in collaboration with @emblebi, @Nvidia and @SeoulNatlUni, we are expanding the database by adding millions of AI-predicted protein complex structures to the AlphaFold Database. To maximise global health impact, we’ve prioritised proteins that are important for understanding human health and disease, including homodimers from 20 of the most studied organisms, including humans, as well as the @WHO’S bacterial priority pathogens list. Read more here: embl.org/news/science-t…

@wtgowers @GoogleDeepMind Agreed @wtgowers. We can definitely explore this - will send you a mail.

@pushmeet @GoogleDeepMind This is very interesting. Is it possible to "interpret" the graphs that AlphaEvolve finds -- e.g. by saying something like "Start with this Paley graph, then modify it by doing X, then Y, etc." ? That would be even better.

Happy to share new progress in AI for Maths @GoogleDeepMind . In extremal combinatorics, AlphaEvolve has helped establish new lower bounds for FIVE classical Ramsey numbers - a problem so challenging that even Erdős commented on its difficulty. Historically, computationally deriving these bounds required bespoke, human-designed search algorithms. For many of these bounds, the best previous results are at least a decade old. AlphaEvolve changes this by acting as a single meta-algorithm that automatically discovers the search procedures needed to find these new bounds. 📷

Our graph constructions for these improved bounds are available at GitHub: github.com/google-researc… Read the full working paper here: arxiv.org/abs/2603.09172