​(view in browser)​

Dear Reader,

Welcome to the July 23rd edition of the Data Science Briefing! This week, we continue celebrating the 6th anniversary of this humble newsletter, and as usual, we have a few surprises in store for you throughout July, starting with the announcement of the next edition of our most successful webinar series: Machine Learning with PyTorch for Developers on Sept 17. Register now so you don't miss out!

This week’s newsletter spotlights some of the most fascinating advances and thought-provoking commentary in AI and data science. In "The Big LLM Architecture Comparison," readers are taken on a deep dive into the latest open-weight language models, such as DeepSeek V3 and Llama 4, examining the architectural innovations, from multi-head latent attention to expert mixtures, that are shaping model performance in 2025. Meanwhile, an exploration from Quanta Magazine reveals just how creative AI has become, highlighting systems that design physics experiments so unconventional that even researchers are stunned by their success in the real world.

In a historic milestone, DeepMind’s Gemini “Deep Think” model stunned the mathematical community by achieving gold-medal status at the 2025 International Mathematical Olympiad, solving university-level math problems entirely in natural language within the contest time frame—marking a world-first for autonomous mathematical reasoning. For those coding on the bleeding edge, reflections on "Coding with LLMs in the summer of 2025" offer a glimpse into the evolving workflow with generative models as indispensable development partners. This issue captures a dynamic portrait of a field racing ahead across algorithms, hardware, and the ever-expanding role of artificial intelligence in science and and engineering.

Fresh from academic, emerging research continues to reshape our understanding of both the natural and digital world. A sweeping global assessment of carnivore populations reveals that 64% of terrestrial carnivores now mainly inhabit regions of heightened human pressure, with Indigenous peoples’ lands providing a last stronghold for nearly a quarter of their ranges—spotlighting both the fragility and importance of these habitats for global biodiversity. On the technological front, billions of Android smartphones are quietly doubling as seismic sensors in 98 countries, now delivering earthquake warnings with effectiveness on par with dedicated scientific instrumentation—an innovation that transforms everyday devices into distributed safety infrastructure and demonstrates unmistakable value for public good.

Meanwhile, new methods like linear scaling causal discovery promise to revolutionize high-dimensional time series analysis, allowing researchers to unveil the deep, hierarchical causal structures in complex data at an unprecedented computational scale. In the realm of digital culture, recent findings on “language bubbles” within online networks expose how community isolation can lock groups into ever more insular linguistic patterns, prompting concern for the erosion of shared discourse. And in a surprising turn for developer productivity, a controlled study reveals that even though open-source developers and experts predicted time savings the adoption of early-2025 AI coding tools actually increased task completion times by 19%, challenging assumptions about the near-term impact of AI in professional software engineering. Across conservation, global risk mitigation, analytics, and social dynamics, these investigations underscore the profound ways science and technology are converging to alter our world.

This week's book is "Behavioral Network Science: Language, Mind, and Society" by T. T. Hills. You can find all the previous book recommendations on our website. In this week's video, we feature an interview of Geoffrey Hinton: I Tried to Warn Them, But We’ve Already Lost Control!

Data shows that the best way for a newsletter to grow is by word of mouth, so if you think one of your friends or colleagues would enjoy this newsletter, go ahead and forward this email to them. This will help us spread the word!

Semper discentes,

The D4S Team

"Behavioral Network Science: Language, Mind, and Society" by T. T. Hills successfully bridges two distinct scientific domains, demonstrating how network analysis can reveal hidden patterns in human behavior. The book tackles an impressive scope of topics, from language evolution and childhood learning to cognitive aging, creativity, and social dynamics, while maintaining remarkable coherence throughout. What sets this work apart is Hills' commitment to practical application, equipping readers with concrete tools, including an introductory guide to network science and accompanying R code that enables hands-on analysis.

This practical approach makes the book uniquely valuable to a diverse audience. Behavioral scientists unfamiliar with network methods will find an accessible entry point, while data scientists can discover rich applications in behavioral research. Hills demonstrates particular skill in addressing contemporary social issues through a network lens, offering fresh perspectives on polarization, echo chambers, and conspiracy theories. The interdisciplinary framework proves especially powerful when examining how individual cognitive processes scale up to shape collective behavior and social structures.

The book's most significant achievement lies in its clarity without oversimplification. Hills effectively conveys complex concepts with precision while maintaining an engaging and accessible tone. This balance makes "Behavioral Network Science" essential reading for anyone seeking to understand how network structures influence human behavior across scales—from individual minds to entire societies.

1. ​The Big LLM Architecture Comparison [magazine.sebastianraschka.com]
2. ​AI Comes Up with Bizarre Physics Experiments. But They Work​ [quantamagazine.org]
3. ​Speeding Up My ZSH Shell [scottspence.com]
4. ​I'm rebelling against the algorithm​ [varunraghu.com]
5. ​State-of-the-Art Multiplatfor​m Matrix Multiplication Kernels​ [burn.dev]
6. ​Why Facts Don’t Change Minds in the Culture Wars—Structure Does [vasily.cc]
7. ​Advanced version of Gemini with Deep Think officially achieves gold-medal standard at the International Mathematical Olympiad​ [deepmind.google]
8. ​Coding with LLMs in the summer of 2025 [antirez.com]

• ​Global scale assessment of the human-induced extinction crisis of terrestrial carnivores (E. J. Torres-Romero, T. M. Eppley, W. J. Ripple, T. M. Newsome, M. Krofel, N. H. Carter, A. Ordiz, T. G. de Oliveira, N. Selva, V. Penteriani)
• ​Global earthquake detection and warning using Android phones(R. M. Allen, A. Barski, M. Berman, R. Bosch, Y. Cho, X. S. Jiang, Y.-L. Lee, S. Malkos, S. M. Mousavi, P. Robertson, B. Spooner, M. Stogaitis, N. Thiruverahan, G. Wimpey)
• ​Measuring how computer science research translates into innovation and development​ (F. Cinus, A. Septiandri, M. Constantinides, D. Quercia)
• ​Linear Scaling Causal Discovery from High-Dimensional Time Series by Dynamical Community Detection​ (M. Allione, V. D. Tatto, A. Laio)
• ​From Signed Networks to Group Graphs​ (T. S. Evans)
• ​Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity​ (J. Becker, N. Rush, E. Barnes, D. Rein)
• ​Language bubbles in online social networks​ (A. Bellina, D. R. LoSardo, E. Brugnoli, F. Saracco, P. Gravino, V. Loreto, G. Di Bona)

Geoffrey Hinton: I Tried to Warn Them, But We’ve Already Lost Control!

​

All the videos of the week are now available in our YouTube playlist.

​Unsubscribe​