Ruha Benjamin on Uneven Tech Distribution & Progressive AI Governance

It’s 2026. A translational scientist starts her day not by searching for papers, but by stepping into a fully revived research context: literature, multi-omics data, hypotheses, simulations, and collaborators — all orchestrated and ready. This is the vision behind Agentic Research. In this blog, we describe what the day-to-day of future scientists and explore how agentic AI transforms scientific work end-to-end: • From context-aware knowledge retrieval across literature, omics, and patents • To constraint-guided hypothesis generation grounded in real human data • To live, collaborative notebooks where reasoning models, simulations, and scientists co-create • To reproducible, slide-ready outputs generated directly from executable analyses Read the full piece and explore what Agentic Research means for the future of R&D: https://hubs.ly/Q03Z5THj0

I think it's a great concept that we don't have to chose between either approach. We can rather combine the accuracy of physics-based methods with the speed of ML-assisted tools

Is the future of drug discovery driven by physics or artificial intelligence? In 2025, the answer became clearer than ever: It’s both. This year, Schrödinger's digital chemistry laboratory reached new heights. From unlocking "undruggable" targets to addressing the industry's most persistent selectivity challenges, we are proud to highlight five landmark publications from 2025 that demonstrate the power of combining the accuracy of physics with the speed of AI.

Assistant Professor in the Division of Physical Therapy Rachel Prusynski will join the Winter 2026 cohort of the University of Washington Data Science & AI Accelerator program through the eScience Institute. The program pairs researchers with data scientists to advance data-intensive projects. Read more: https://ow.ly/cqoY50XQ75I

🔍 The core value of scientific visualization is never "beautifying data" but making complex laws intuitively perceptible! Just like nanoscale fluorescence imaging reveals neural structures, standardized color matching makes data transmission more equitable, and tool upgrades make scientific expression more efficient. Every qualified scientific figure is a balance between science and aesthetics. What percentage do you think visualization contributes to scientific communication? #Research_Visualization_Value #Academic_Dissemination #Scientific_Storytelling

Bioimaging enables researchers to visualise and understand biological life, driving scientific innovation and discovery. Our new report draws on expert insights to set out a roadmap for progress in the next 10 years. From improving data standards and integrating AI, to ensuring equitable access – advancing the field requires a collective effort from researchers, industry and funders. Find out more ⤵️ https://lnkd.in/e8y_ez4z

Surrogate Model Achieves Sub-Angstrom Accuracy in Molecular Dynamics Simulations A new computational framework accurately predicts how atoms move in metallic systems, such as aluminum, by learning the underlying rules of motion and bypassing the need for traditional, computationally intensive calculations. #quantum #quantumcomputing #technology https://lnkd.in/eXJcEi5J

Researchers from the Indian Institute of Technology Delhi, along with collaborators from Denmark and Germany, have turned it into reality, as detailed in their recent Nature Communications publication titled "Evaluating large language model agents for automation of atomic force microscopy." #IITDelhi #NatureCommunications #AIInScience #LargeLanguageModels #ScientificAutomation https://lnkd.in/gaacXp5f

On physics' foundations, looks a real breakthrough. "New research from MIT reveals a breakthrough: wildly different AI models for molecules, proteins, and materials are all independently converging on the same internal representation of matter. This suggests we aren't just building simulators; we’re uncovering a shared, physics-grounded Latent Reality. We might finally be approaching a true, universal Scientific Foundation Model that understands the building blocks of everything." Original link in the comments. Michael Erlihson PhD Mikhail Gorelkin Charles H. Martin, PhD

We’ve just released AceFF™-2, our most advanced machine-learning force field to date. The goal? To give drug discovery teams the precision of Quantum Mechanics at the speed of classical MD. Why it matters: 🔹 Faster lead optimization. 🔹 More accurate binding predictions. 🔹 A massive leap forward for AI-driven drug discovery. Read the blog post to see how we’re bridging the gap: 🔗 https://lnkd.in/excitprj