How Simple Rules Enable Microorganisms—and Future Nanobots—to Swim Without a Brain

Hello and welcome to our May 20th edition. The STEAM Digest is a curated newsletter that brings you the latest in science, technology, engineering, arts, and mathematics.

In today’s edition:

• Science - Researchers Create World's Fastest Petahertz Transistor Using Light and Graphene, and more.

• Materials - Self-Adaptive Glasses Take Shape: Study Reveals How Activity Alters Material Behavior and Memory.

• Biotechnology & Biomedical Technology - How Simple Rules Enable Microorganisms—and Future Nanobots—to Swim Without a Brain, and more.

• Engineering & Technology - Shape-Shifting Dimpled Skin Could Revolutionize Underwater and Aerial Vehicle Design, and more.

• Robotics, AI, Hardware, Software, Gadgets - WildFusion Enables Robots to Navigate Forests Using Sight, Sound, and Touch - Like Humans, and more.

• Astronomy, Space, & Astrobiology - New Study Suggests Mysterious Martian Slope Streaks Are Dry, Not Evidence of Water, and more.

• Health & Medicine - Engineered T Cells Offer First Potential Treatment for Celiac Disease, and more.

• Neuroscience - Duke’s Experimental Drug Offers Potent Pain Relief Without Opioid Side Effects.

• Environment & Earth Sciences - Warming Arctic Seas Fuel Seaweed Growth, Raising Risk of Oxygen Loss in Fjords, and more.

• Nature & Ecology - Scientists Decode Genome of Strange Worm That Reproduces Through Detachable Tails, and more.

• Other Sciences & The Arts - Choice Engineering: Math-Driven Models Outperform Psychology in Shaping Decisions.

Until Tomorrow,

~The STEAM Digest

If you would like to share The STEAM Digest newsletter with others, please use the following link: https://thesteamdigest.beehiiv.com/subscribe

SCIENCE

Researchers Create World's Fastest Petahertz Transistor Using Light and Graphene: Scientists have developed a graphene-based quantum transistor capable of operating at petahertz speeds—more than 1,000 times faster than current processors. The team used ultrafast laser pulses lasting just 638 attoseconds to manipulate electrons through quantum tunneling, allowing them to bypass barriers almost instantly. This effect was observed using a modified graphene-silicon phototransistor. The breakthrough represents a major leap toward light-based computing. Unlike many experimental technologies, the transistor operates under normal conditions, making it viable for future real-world applications. Researchers hope this advancement will pave the way for ultrafast computing in fields like AI, medicine, and space exploration. They are now working to commercialize the technology with the help of Tech Launch Arizona.

Cutting Onions with Sharp, Slow Blades Reduces Eye Irritation, Cornell Study Finds: A multidisciplinary team at Cornell University has identified key factors that influence how much irritating spray is released when cutting onions. The researchers found that sharper knives and slower cutting speeds significantly reduce the release of syn-propanethial-S-oxide—the compound responsible for eye irritation. Using a specially designed guillotine, painted onion chunks, and high-resolution imaging, the team discovered that dull knives compress onion layers, causing them to spring back and spray juice into the air. Faster cuts also produce more mist, increasing irritation. Their conclusion: to minimize eye sting, use the sharpest knife available and cut onions slowly.

MATERIALS

Self-Adaptive Glasses Take Shape: Study Reveals How Activity Alters Material Behavior and Memory: Researchers have shown through computer simulations that active glasses—disordered materials with self-powered components—can adapt their mechanical properties in response to internal motion. The study reveals that activity within poorly annealed glasses can trigger self-annealing, transforming ductile materials into brittle ones by helping the system explore lower-energy configurations. These active glasses also exhibit memory effects and behavior similar to passive glasses under cyclic shear, offering insights into biological tissue dynamics, wound healing, and even learning processes. The findings could inform the design of new metamaterials with tunable properties and bridge soft matter physics with biophysics. While this method doesn’t yet match advanced annealing techniques like Swap Monte Carlo, it opens new research avenues into how materials remember, adapt, and evolve.

BIOTECHNOLOGY & BIOMEDICAL TECHNOLOGY

How Simple Rules Enable Microorganisms—and Future Nanobots—to Swim Without a Brain: A new study demonstrates that goal-directed swimming is possible without a central control system, like a brain. Using computer simulations, the team modeled microorganisms as chains of interconnected beads, each following basic local rules and equipped with a tiny neural network. Despite no awareness of the whole structure, the beads collectively produced efficient, coordinated movement through viscous fluids. The findings explain how single-celled organisms like bacteria move and suggest that nanobots could be built to perform complex tasks—such as targeted drug delivery or environmental cleanup—using only simple decentralized programming.

AI Detects Swarming Bacteria from a Single Image—A Potential Diagnostic Breakthrough: Researchers from UCLA and Albert Einstein College of Medicine have developed an AI method that accurately detects bacterial swarming—a collective movement linked to both infections and healing—from a single blurry microscope image. Swarming, distinct from swimming, plays a dual role in diseases like UTIs and IBD, making it a promising diagnostic biomarker. The deep learning model was trained on thousands of time-encoded images, achieving 97.44% sensitivity and 100% specificity, and it even generalized well to new bacterial strains. Unlike traditional video-based methods, this automated, image-only approach is faster, more scalable, and suitable for low-resource settings. The team envisions future applications in smartphone-based diagnostics and on-site bacterial monitoring.

ENGINEERING & TECHNOLOGY

Shape-Shifting Dimpled Skin Could Revolutionize Underwater and Aerial Vehicle Design: Inspired by golf ball dimples, University of Michigan researchers have developed a prototype sphere with a programmable surface that reduces drag and enables precise movement. The device uses a vacuum-controlled latex skin to create adjustable dimples, dynamically responding to airflow to cut drag by up to 50%. At higher speeds, shallower dimples are more effective; deeper ones work best at lower speeds. The surface can also generate lift for steering by dimpling only one side, enabling controlled movement without fins or rudders. This breakthrough could revolutionize the design of compact, energy-efficient underwater and aerial vehicles.

SonoTextiles: Researchers Unveils Smart Fabrics That Use Sound to Sense Motion and Breathing: Researchers have developed SonoTextiles, smart fabrics that use acoustic waves through glass fibers—instead of electronics—to detect motion, pressure, and breathing. Unlike conventional smart textiles, this approach is lighter, more breathable, washable, energy-efficient, and cost-effective. Each fiber carries ultrasonic waves at distinct frequencies, allowing accurate real-time tracking with minimal computing power. Applications range from monitoring asthma patients’ breathing and sports performance to sign language translation and posture correction. While still in the lab stage, the team is working to make the system more robust, including replacing fragile glass fibers with durable metals.

ROBOTICS, AI, HARDWARE, SOFTWARE, GADGETS

WildFusion Enables Robots to Navigate Forests Using Sight, Sound, and Touch - Like Humans: Duke University researchers have developed WildFusion, a robotic navigation system that combines vision, touch, vibration, and balance to help robots move through complex environments like forests. Traditional robots struggle in such settings due to reliance on cameras or LiDAR alone. WildFusion uses added tactile sensors, contact microphones, and inertial data to sense terrain changes like mud or leaves, enabling more confident and adaptive movement. A deep learning model fuses all sensory inputs to form a continuous 3D understanding of the environment, even when visual data is unclear. Tested successfully in North Carolina's Eno River State Park, WildFusion helped a quadruped robot identify safe paths. The system could enhance disaster response, remote inspections, and autonomous exploration. Future upgrades may include thermal and humidity sensors.

AI Model Optimizes Global Land Use for Climate and Sustainability Goals: Researchers at UT Austin and Cognizant AI Labs have created an AI system that analyzes 175 years of global land use and carbon data to suggest optimal policies for boosting carbon storage with minimal economic impact. The AI uses an evolutionary approach to test thousands of land-use scenarios, revealing that targeted, location-specific changes—like converting cropland to forest—are more effective than blanket solutions. The tool helps decision-makers balance environmental and economic trade-offs and could also be used to tackle other global challenges like food security and public health..

GPT-4 Found More Persuasive Than Humans in Targeted Online Debates: A has study found that GPT-4 was more persuasive than humans in 64% of online debates when given personal information about its opponent. Researchers matched 900 U.S. participants to debate sociopolitical issues with either a human or GPT-4, some with access to demographic data like age, gender, and political affiliation. GPT-4 significantly outperformed human debaters only when it used this personalized information to tailor its arguments. Without it, its persuasiveness matched that of humans. The findings raise concerns about AI's ability to influence opinions and suggest a need for safeguards in online interactions. Limitations include the study’s structured debate format and time constraints.

ASTRONOMY, SPACE, & ASTROLOBIOLOGY

New Study Suggests Mysterious Martian Slope Streaks Are Dry, Not Evidence of Water: A new study challenges the long-held theory that dark streaks on Martian slopes—known as slope streaks and recurring slope lineae (RSL)—are formed by liquid water. Using machine learning to analyze over 86,000 satellite images, the team mapped more than 500,000 streaks and found no correlation with signs of water, such as high humidity or temperature. Instead, the features were strongly associated with wind activity and dust movement, suggesting they form when fine dust slides down slopes—possibly triggered by impacts or dust devils. The findings reduce the likelihood that these streaks represent habitable environments and lessen concerns over contaminating such sites with Earth-based microbes during future Mars missions.

Evolving Dark Matter: A New Twist on the Hubble Tension Mystery: A new study proposes that dark matter may evolve over time, offering a potential solution to the long-standing Hubble tension—the mismatch between measurements of the universe's expansion rate in the early and modern cosmos. While evolving dark energy has been widely explored, evolving dark matter has received little attention until now. The authors suggest a model in which about 15% of cold dark matter has an oscillating equation of state, similar to how neutrinos oscillate in mass. This mixed model fits current observations better than those involving evolving dark energy alone. Though still a theoretical “toy model”, the research opens a new path in cosmology, hinting that tweaking—not discarding—dark matter may help resolve one of modern physics' biggest puzzles.

NASA and Researchers Develop Pyastrobee Simulator to Enhance Space Robot Cargo Handling: Astrobee, NASA’s free-flying robotic system aboard the International Space Station (ISS), assists astronauts with routine tasks but struggles with manipulating deformable objects like cargo bags. To address this, researchers have developed Pyastrobee, a Python-based simulation and control toolkit. Presented on the arXiv preprint server, Pyastrobee realistically models the ISS environment and deformable cargo using the Bullet physics engine. It integrates with reinforcement learning frameworks like Gymnasium and Stable Baselines to train Astrobee in handling cargo more efficiently. The open-source platform enables testing of advanced control strategies, including model-predictive control (MPC), and aims to make space robot operations more autonomous—especially valuable for future uncrewed missions.

HEALTH & MEDICINE

Engineered T Cells Offer First Potential Treatment for Celiac Disease: Researchers have developed a promising cell therapy for celiac disease, currently only managed through a gluten-free diet. The study adapts CAR T cell technology to engineer regulatory T cells (Tregs) that suppress harmful immune responses to gluten. In an animal model, these Tregs prevented inflammation and stopped damaging T cells from attacking the intestine. This marks the first potential therapeutic approach to restore gluten tolerance. Further research is needed to test effectiveness in humans.

Acetate and Gut Bacteria Combination Found to Reduce Obesity in Mice: Researchers have discovered that a supplement delivering acetate to the gut—called AceCel—reduces fat and liver mass in mice, but only when Bacteroides bacteria are present. The study showed that AceCel promotes fat burning and limits sugar absorption by enhancing carbohydrate fermentation in the gut. This metabolic shift mimics the effects of fasting or a keto diet, encouraging weight loss without muscle loss. In mice lacking gut bacteria, the effect disappeared, confirming the essential role of Bacteroides. The team now aims to test AceCel's safety and effectiveness in humans as a potential ingredient in anti-obesity functional foods.

ALDH4A1 Identified as Key Tumor-Suppressive Component of Mitochondrial Pyruvate Carrier Complex: Researchers have discovered that ALDH4A1, a mitochondrial enzyme, is a third structural component of the mitochondrial pyruvate carrier (MPC) complex, alongside MPC1 and MPC2. This trimeric complex is crucial for importing pyruvate into mitochondria, a key step in cellular energy production. The study shows that ALDH4A1 stabilizes the MPC complex, supports oxidative metabolism, and suppresses cancer growth. Loss of ALDH4A1 promoted cancer-like behavior in cells and accelerated tumor growth in mice, while its overexpression reduced tumor size. The MPC inhibitor UK5099 disrupted ALDH4A1’s interaction with the complex, highlighting a potential therapeutic target. These findings expand ALDH4A1’s known role beyond proline metabolism and suggest its importance in cancer-related metabolic regulation. Targeting the ALDH4A1-MPC interface may offer new strategies for cancer treatment.

NEUROSCIENCE

Duke’s Experimental Drug Offers Potent Pain Relief Without Opioid Side Effects:
Researchers at Duke University School of Medicine have developed a promising non-opioid painkiller, SBI-810, which relieves pain without the side effects or addiction risks associated with opioids. The study shows SBI-810 targets the neurotensin receptor 1 and activates a specific pain-relief pathway—β-arrestin-2—avoiding the euphoric and harmful pathways triggered by opioids. In mouse models, SBI-810 outperformed both opioids and gabapentin in treating pain from surgery, fractures, and nerve damage, without causing sedation, tolerance, or constipation. It also enhanced the effect of opioids when used in combination, allowing for lower doses. The drug could be especially valuable for patients with chronic or post-surgical pain. Human trials are the next step, and multiple patents have already been secured.

ENVIRONMENT & EARTH SCIENCES

Warming Arctic Seas Fuel Seaweed Growth, Raising Risk of Oxygen Loss in Fjords: A new study reveals that rising sea temperatures and shrinking sea ice are driving seaweed expansion in Arctic fjords like Svalbard’s Kongsfjorden. The research team analyzed seaweed and sediment samples, finding that dense seaweed growth may trigger low-oxygen zones due to microbial activity. While different seaweed types absorb carbon differently, they share similar lipid profiles, indicating adaptation to the Arctic environment. Sediment analysis showed that inner fjord areas receive carbon from varied sources, while outer fjords are more influenced by seaweed. The findings suggest that unchecked seaweed expansion could destabilize Arctic coastal ecosystems by altering carbon cycling and reducing oxygen levels.

Corroded Underwater Bombs Pose Hidden Danger—Acoustic Research Aims to Improve Detection: Over 400 underwater sites in the U.S. may contain unexploded ordnance (UXO), raising safety concerns, especially in shallow waters. Connor Hodges, a Ph.D. student at the University of Texas at Austin, is studying how corrosion and biofouling alter the acoustic signatures of aging UXOs, making them harder to detect using sonar. At the Acoustical Society of America meeting, Hodges presented findings from tests on AN-Mk 23 practice bombs buried for 80 years, showing that decay weakens sonar signals and risks misclassification. As former military zones transition to public use, improved acoustic detection methods could play a key role in civilian demining and safety efforts. Hodges’ research aims to develop better tools to find and safely recover these dangerous remnants.

Global Glaciers Won’t Recover for Centuries After Climate Overshoot, Study Finds: New research shows that global mountain glaciers will not recover for centuries, even if global temperatures are eventually brought back down to 1.5°C after exceeding it. The study modeled glacier changes through the year 2500 under a "climate overshoot" scenario where temperatures rise to 3°C before declining. The findings show that such a scenario would cause glaciers to lose up to 16% more mass than if the 1.5°C threshold had never been crossed, further accelerating sea-level rise. Even with long-term cooling, glacier regrowth would take centuries or longer, with serious implications for freshwater availability in regions like the Alps, Andes, and Himalayas. Researchers emphasize that overshooting climate goals causes irreversible glacier loss, underscoring the urgent need for rapid emission reductions.

NATURE & ECOLOGY

Scientists Decode Genome of Strange Worm That Reproduces Through Detachable Tails: Researchers have mapped the gene activity of Ramisyllis kingghidorahi, a rare branching worm that reproduces by growing detachable tails called stolons. These stolons act as independent reproductive units that swim off to mate. The study found the greatest gene activity in the stolons, not the head, especially in genes linked to eye development. This helps explain how stolons prepare for life outside the body. Researchers also discovered signs of partial genome duplication, which may support the worm’s unusual form. The findings offer new insight into bizarre reproductive strategies and body structures in marine life.

Mice Use Chemical Cues to Judge Social Rank and Decide How to Behave, Study Finds: Researchers have discovered that mice use chemical signals—both airborne odors and contact-based cues—to detect and compare the social rank of unfamiliar mice, guiding their own behavior in social situations. The study shows that mice don’t rely on physical traits or fixed behaviors, but instead make brain-based decisions based on inferred relative dominance. Disabling both their olfactory and vomeronasal systems prevented rank detection, confirming that both sensory pathways are essential. This mechanism mirrors how humans gauge social status using visual and social cues. The next step is to identify the brain regions responsible for processing this ranking information and triggering submissive or aggressive responses.

OTHER SCIENCES & THE ARTS

Choice Engineering: Math-Driven Models Outperform Psychology in Shaping Decisions: A new study introduces choice engineering, a powerful method for influencing decisions using mathematical models instead of psychological intuition. The study found that computational strategies outperformed traditional behavior-based approaches in guiding human choices. In a large-scale experiment with over 3,000 participants, the top-performing method—called CATIE—used data-driven modeling to predict and influence decision-making more accurately than intuition or standard machine learning. Unlike choice architecture, which relies on psychological nudges, choice engineering uses precise, scalable models to optimize behavioral interventions. The approach could transform fields like education, health, and policy—but also raises ethical questions about how these tools are applied.