The appearance of a hot sauce or pepper doesn’t reveal whether it’s mild or likely to scorch someone’s taste buds. So, researchers made an artificial tongue to quickly detect spiciness. Inspired by milk’s casein proteins, which bind to capsaicin and relieve the burn of spicy foods, the researchers incorporated milk powder into a gel sensor. The prototype, reported in ACS Sensors, detected capsaicin and pungent-flavored compounds (like those behind garlic’s zing) in various foods.

Crystal structure prediction (CSP) of organic molecules is a critical task, especially in pharmaceuticals and materials science. However, conventional methods are computationally intensive and time-consuming. Now, researchers from Japan have developed a new workflow: SPaDe-CSP that accelerates CSP by machine learning-based prediction of most probable space groups and crystal densities and employing an efficient neural network potential for structure refinement. It achieved faster and more reliable CSP than conventional methods.

In nature, living systems effortlessly sense, move, and adapt to changing environments. Replicating such dynamic behavior in artificial materials has long challenged scientists. A recent study introduces supramolecular robotics—a molecular design strategy that enables soft materials to exhibit autonomous motion, reversible transformations, and tissue-like organization. This innovation marks a key step toward creating programmable, life-like systems that blur the line between chemistry and robotics.