David Baker spoke with WIRED en Español about where biological design is headed next and the great potential and risks of taking biology outside the bounds of the natural world. David Baker, a pioneer of protein design, is using AI to go where nature hasn't. The natural world as we know it represents only a fraction of what might exist. Based on this idea, AI BioDesign was born: a scientific project that combines artificial intelligence and large-scale laboratory experiments to design and test new molecules and biological functions that are not found in nature but are physically and chemically possible. In doing so, the researchers are aiming to create databases, models, and tools that could serve as "seeds" for developing the medicines and technologies of the future.

Diffusion-based models have recently enabled the generation of realistic and diverse protein structures, yet they remain limited in their ability to steer outcomes toward specific functional or biochemical objectives, such as binding affinity or sequence composition. Here we extend the Feynman-Kac (FK) steering framework, an inference-time control approach, to diffusion-based protein design. By coupling FK steering with structure generation, the method guides sampling toward desirable structural or energetic features while maintaining the diversity of the underlying diffusion process. To enable simultaneous generation of both sequence and structure properties, rewards are computed on models refined through ProteinMPNN and all-atom relaxation. Applied to binder design, FK steering consistently improves predicted interface energetics across diverse targets with minimal computational overhead. More broadly, this work demonstrates that inference-time FK control generalizes diffusion-based protein design to arbitrary, non-differentiable, and reward-agnostic objectives, providing a unified and model-independent framework for guided molecular generation.

Large Language Models (LLMs) have demonstrated considerable success in open-book question answering (QA), where the task requires generating answers grounded in a provided external context. A critical challenge in open-book QA is to ensure that model responses are based on the provided context rather than its parametric knowledge, which can be outdated, incomplete, or incorrect. Existing evaluation methods, primarily based on the LLM-as-a-judge approach, face significant limitations, including biases, scalability issues, and dependence on costly external systems. To address these challenges, we propose a novel metric that contrasts the perplexity of the model response under two conditions: when the context is provided and when it is not. The resulting score quantifies the extent to which the model's answer relies on the provided context. The validity of this metric is demonstrated through a series of experiments that show its effectiveness in identifying whether a given answer is grounded in the provided context. Unlike existing approaches, this metric is computationally efficient, interpretable, and adaptable to various use cases, offering a scalable and practical solution to assess context utilization in open-book QA systems.

A day after recognizing former Google vice president and engineering fellow Geoffrey Hinton for his contributions to the field of physics, the Royal Swedish Academy of Sciences has honored a pair of current Google employees. On Wednesday, DeepMind CEO Demis Hassabis and senior research scientist John Jumper won half of the 2024 Nobel Prize in Chemistry, with the other half going to David Baker, a professor at the University of Washington. Baker, Hassabis and Jumper all advanced our understanding of those essential building blocks of life that are responsible for functions both inside and outside the human body. The Nobel Committee cited Baker's seminal work in computational protein design. Since 2003, Baker and his research team have been using amino acids and computers to design entirely new proteins.

Google DeepMind founder Demis Hassabis has won a joint Nobel Prize for Chemistry for using artificial intelligence to predict the structures of proteins. Hassabis shares half the prize with John M. Jumper, a director at Google DeepMind, while the other half has been awarded to David Baker, a professor in biochemistry at the University of Washington for his work on computational protein design. The potential impact of this research is enormous. Proteins are fundamental to life, but understanding what they do involves figuring out their structure--a very hard puzzle that once took months or years to crack for each type of protein. By cutting down the time it takes to predict a protein's structure, computational tools such as those developed by this year's award winners are helping scientists gain a greater understanding of how proteins work and opening up new avenues of research and drug development.

The 2024 Nobel Prize in Chemistry has been awarded to a trio of scientists for their breaththrough work into protein structures. London-born Demis Hassabis, CEO of British AI firm, DeepMind, is one of the three given the prize, along with his colleague John M. Jumper and American biochemist David Baker. Together, they cracked the code for proteins' amazing structures, which had previously been much of a mystery. 'One of the discoveries being recognised this year concerns the construction of spectacular proteins,' said Heiner Linke, Chair of the Nobel Committee for Chemistry. 'The other is about fulfilling a 50-year-old dream – predicting protein structures from their amino acid sequences.

The 2024 Nobel prize in chemistry has been awarded to David Baker, Demis Hassabis and John Jumper for their work on understanding the structure of proteins, which play vital roles in all living organisms. Hassabis and Jumper, of Google DeepMind, developed an artificial intelligence that predicts the structure of proteins. Baker, at the University of Washington in Seattle, has been recognised for his work on designing new proteins. Proteins are the molecules that make life happen. All of the key machinery of life is made of proteins, from the muscles that power us and the molecules that read and copy DNA to the antibodies that protect us from infections.

The world of AI-powered drug discovery keeps expanding as the capabilities of machine learning grow. One approach that seemed unthinkable just a few years ago is simulating the complicated interplays of two interlocking molecules -- but that's exactly what drug designers need to know about, and exactly what Charm Therapeutics aims to do with its DragonFold platform. Proteins do just about everything worth doing in your body, and are the most frequent targets for drugs. And in order to create an effect, you must first understand that target, specifically how the chain of amino acids making up the protein "folds" under different circumstances. In the recent past this was often done with complex, time-consuming X-ray crystallography, but it has recently been shown that machine learning models like AlphaFold and RoseTTAFold are capable of producing results just as good but in seconds rather than weeks or months.