Weekly Piece of Future #133
From Brain Age Reversal to Neuromorphic AI and Artificial Photosynthesis
Hey there, fellow future-addicts!
Welcome to this week's edition of Rushing Robotics, your weekly update on the most incredible breakthroughs in science and technology. Innovation is moving faster than ever. This week, we explore breakthroughs that feel like science fiction brought to life—from reversing brain aging to AI-driven robots and lab-grown human skin.
🤯 Mind-Blowing
Prepare to have your perception of what's possible completely reshaped. This week, scientists are challenging long-held assumptions and redefining the boundaries of biology and technology. We'll explore a newly discovered protein that can reverse brain aging, a breakthrough in artificial photosynthesis that could lead to carbon-neutral fuels, and the unveiling of "Wukong," the world's largest brain-inspired supercomputer, designed to unlock the secrets of the human mind and pave the way for true artificial general intelligence.
🔊 Industry Insights & Updates
From the factory floor to the data center, AI is no longer a concept—it's a physical reality. In this section, we'll cover the latest moves from industry giants like NVIDIA as they transition generative AI from the cloud to embedded robotics. We'll also examine how Cornell researchers are revolutionizing 3D printing for superconductors, how Japan is pioneering new renewable energy sources, and how a new generation of robots is making on-demand delivery profitable.
🧬 BioTech
The future of medicine is here, and it's built on a foundation of groundbreaking biological research. This week, we'll delve into a new understanding of DNA mechanics that could transform genetic studies, the world's first pig-to-human lung transplant, and a revolutionary method for early disease detection that uses light and matter to reveal molecular secrets. These stories showcase how biotech is pushing the limits of what's possible for human health.
💡 Products/Tools of the Week
Looking for the next tool to revolutionize your workflow? This section is your go-to guide for the most innovative products and platforms on the market. From a suite of medical AI APIs designed to streamline healthcare, to a full-stack application generator that turns your ideas into functional web apps, and AI-powered platforms for everything from financial planning to building intelligent agents without writing a single line of code.
🎥 Video Section
From ETH Zürich’s robot learning badminton to Unitree’s record-breaking robotic feats, this week’s videos showcase AI and robotics in action. Watch machines move, learn, and compete in ways that blur the line between science and spectacle.
Every story in this newsletter is a glimpse of how technology and science are rewriting the rules. The pace of change is breathtaking, and we’re only getting started. Stay hungry, stay futurish!
🤯 Mind-Blowing
Scientists have found a protein that reverses brain aging. They have identified a surprising culprit behind age-related cognitive decline: a protein known as FTL1. In a recent study involving mice, elevated levels of FTL1 were found to cause a variety of impairments, including memory loss, weakened connections between brain cells, and reduced cellular efficiency. The researchers were able to demonstrate a causal link, as artificially increasing FTL1 in young mice led to premature brain aging. Conversely, when they inhibited the protein in older mice, the animals experienced a reversal of these impairments, with a restoration of youthful brain function and sharp memory. This discovery indicates that FTL1 could function as a master switch for aging in the brain, suggesting that targeting it could one day provide a means to reverse cognitive decline rather than merely slow its progression.
China has unveiled Darwin Monkey, also called “Wukong,” the world’s largest brain-inspired supercomputer, designed to replicate the architecture of a monkey’s brain. The system incorporates over 2 billion artificial neurons and more than 100 billion synapses, roughly equivalent to the neural network of a macaque. Researchers envision Darwin Monkey as a powerful simulation tool for neuroscientists, enabling them to study brain function in ways that were previously impossible, while also serving as a stepping stone toward artificial general intelligence (AGI)—AI capable of human-like reasoning, learning, and problem solving. Unlike conventional artificial neural networks, which rely on continuously changing binary signals and classical computing, Darwin Monkey operates on spiking neural networks (SNNs). These networks emulate the way mammalian neurons transmit signals through bursts of electrical activity, or spikes, allowing the system to process and transmit information in a manner that closely resembles a real brain. This neuromorphic approach could open new frontiers for AI research, from cognitive modeling to advanced machine learning architectures.
A fully functional human skin model has been created in the lab, complete with blood vessels, pigmentation, and multiple tissue layers, marking a world-first breakthrough with major implications for skin grafting, wound repair, and chronic disease treatments. Scientists at the University of Queensland’s Frazer Institute developed this advanced skin using stem cells, achieving features like capillaries, hair follicles, and immune cells. After six years of research, the team believes this innovation will transform clinical practice. Dr Abbas Shafiee noted its potential for skin transplants and regenerative medicine, while Professor Kiarash Khosrotehrani highlighted its promise in tackling inflammatory disorders such as psoriasis, atopic dermatitis, scleroderma, and genetic conditions.
Researchers from Shibaura Institute of Technology, Waseda University, and Fujitsu have developed a quantum computing-based method to enable robots to move more smoothly and efficiently. Robotic movement relies on solving inverse kinematics, the calculation of joint angles needed for a robot’s hand or foot to reach a target position. For humanoid robots, the vast number of possible joint configurations makes this process extremely complex. Traditional computers rely on trial-and-error computations, consuming significant time and resources. The new approach leverages qubits to encode the position and orientation of each robot component. Crucially, the method employs quantum entanglement, linking qubits so that the movement of one automatically influences the others—mirroring how connected joints affect one another in real robots. The researchers also adopted a hybrid computing strategy, where quantum circuits handle forward kinematics—determining the end positions given specific joint angles—while classical computers continue to manage inverse kinematics, achieving an efficient balance between speed and precision.
Artificial photosynthesis may be advancing more rapidly than anticipated. Researchers at the University of Basel have engineered a molecule capable of storing four charges derived from light energy, representing a significant step toward the development of carbon-neutral fuels. In biological systems, photosynthesis enables plants to capture solar energy, converting atmospheric carbon dioxide into carbohydrate molecules that serve as energy reserves. Animals and humans metabolize these carbohydrates, releasing the stored energy while returning carbon dioxide to the atmosphere, thereby maintaining the fundamental cycle that sustains life on Earth. Scientists aim to replicate this process in controlled environments. Harnessing sunlight to synthesize high-energy compounds such as hydrogen, methanol, or synthetic hydrocarbons could provide fuels that, when combusted, do not introduce additional carbon dioxide into the atmosphere, rendering them effectively carbon-neutral.
🔊 Industry Insights & Updates
Generative AI is transitioning from cloud-based systems to embedded robotics as NVIDIA introduces its Jetson Thor computing platform for real-time robotic intelligence. Announced on Monday, the company confirmed the general availability of the Jetson AGX Thor developer kit and production modules.
The Jetson Thor system delivers up to 2,070 FP4 teraflops of AI performance and 128 GB of memory while maintaining operation within a 130-watt power envelope. Compared to its predecessor, Jetson Orin, the new platform provides 7.5× greater AI compute and 3.5× improved energy efficiency.
Early adopters include Amazon Robotics, Boston Dynamics, and Caterpillar, with additional trials underway by John Deere, Meta, OpenAI, and Medtronic to accelerate development of what NVIDIA terms “physical AI”—robots capable of perception, reasoning, and autonomous action in real-world environments.
Powered by the Blackwell GPU architecture, Jetson Thor supports concurrent execution of multiple AI models, including vision-language-action frameworks such as NVIDIA’s Isaac GR00T, as well as large-scale language and vision models. This capability enables robots to interact with humans and function in complex environments without relying solely on cloud-based processing.
Cornell researchers have developed an innovative one-step 3D printing process for producing superconductors with unprecedented performance characteristics. This streamlined fabrication method has the potential to significantly advance technologies such as MRI magnets, quantum computing hardware, and other superconducting applications. The achievement builds upon nearly a decade of work led by Ulrich Wiesner, the Spencer T. Olin Professor in Cornell’s Department of Materials Science and Engineering. The latest study introduces a technique that employs an ink composed of copolymers and inorganic nanoparticles, which self-assemble during the 3D printing process. Subsequent heat treatments transform these printed structures into porous crystalline superconductors. This “one-pot” approach eliminates many of the intermediate steps associated with conventional fabrication methods. Such simplification may prove highly beneficial for quantum materials research and the development of next-generation superconducting devices.
Japan has commissioned its first osmotic power plant in Fukuoka, making it the second such facility in the world and the first in Asia. According to the Fukuoka District Waterworks Agency, the plant commenced operations on August 5 and is projected to generate approximately 880,000 kilowatt-hours annually. This energy output will primarily supply a desalination plant serving Fukuoka and surrounding areas. The agency has characterized osmotic power as a “next-generation renewable energy source” that operates independently of weather conditions and diurnal cycles, while producing zero carbon emissions. Positioned as a key development in achieving continuous clean energy, the Fukuoka installation is anticipated to deliver sufficient electricity to power about 220 households in Japan.
Robomart, a company specializing in self-driving delivery robots, is introducing its new robot, the Robomart RM5, with the aim of making on-demand delivery profitable. The Los Angeles-based startup announced the patented level-four autonomous vehicle on Monday. The RM5 can carry up to 500 pounds and is equipped with 10 individual lockers to hold customer orders. This design allows for batch ordering, enabling a single robot to complete multiple deliveries simultaneously. According to Robomart co-founder and CEO Ali Ahmed, this innovation reduces the cost of a delivery by up to 70%, a critical factor when compared to the high cost of paying a human driver.
🧬 BioTech
For years, scientists assumed that DNA subjected to mechanical stress would form knots, creating disordered and tangled configurations. However, recent research from the University of Cambridge challenges this notion, revealing that stressed DNA instead coils into highly organized, spring-like formations known as plectonemes. This discovery fundamentally alters long-standing models of DNA mechanics. The findings emerged from experiments involving nanopores—nanoscale openings just wide enough to accommodate a single DNA strand. In these tests, DNA was immersed in a saline, alkaline solution and subjected to both an applied voltage and fluid flow, driving it through the nanopore. These forces induced rotation, generating torque sufficient to twist the DNA molecule into its coiled state. The implications of this study are significant for both biological science and technological applications. Within living cells, DNA is frequently exposed to torsional stress, whether during chromosomal compaction in the nucleus or through enzymatic activity involving cutting, twisting, and rejoining of strands. If DNA indeed forms plectonemic structures under such conditions, these configurations could influence gene accessibility, replication dynamics, and regulatory mechanisms.
For the first time in medicine, a pig lung was transplanted into a brain-dead human and worked for nine days. The surgery was performed by doctors at Guangzhou Medical University in China. The recipient, a 39-year-old man with a brain hemorrhage, received a left lung from a genetically modified Chinese Bama Xiang pig. At first, the lung functioned properly, providing oxygen and removing carbon dioxide. But within 24 hours, it started to show damage and fluid buildup. Lung xenotransplants are especially difficult because lungs are highly sensitive to what they breathe in, making them prone to rejection. After a few days, the patient’s immune system began attacking the lung, causing severe damage by day six. The team ended the experiment on the ninth day due to antibody-mediated rejection.
Researchers have redefined the fundamentals of light–matter interaction to enable earlier disease detection than previously possible. A team at Johns Hopkins University has introduced an innovative approach for observing molecular vibrations by creating hybrid states between light and molecules, amplifying even the weakest vibrational signals. This breakthrough, led by Professor Ishan Barman of the Department of Mechanical Engineering, holds significant potential for early diagnosis of conditions such as infections, metabolic disorders, and cancer. Molecular vibrations—minute, distinct atomic movements within molecules—serve as highly specific chemical “fingerprints,” offering unprecedented clarity in molecular identification. The implications extend beyond healthcare. In pharmaceutical manufacturing, this technology could facilitate real-time monitoring of intricate chemical reactions, ensuring process consistency and product safety. Additionally, environmental applications include the detection of pollutants and hazardous compounds at trace concentrations with unmatched precision.
💡Products/tools of the week
The MedGemma and MedSigLIP APIs constitute an advanced suite of medical AI services designed to equip healthcare professionals and developers with cutting-edge tools for analyzing clinical data and generating medical content. These APIs integrate sophisticated natural language processing and computer vision technologies to support medical text generation, clinical question answering, and comprehensive multimodal analysis of textual and imaging data. By leveraging these APIs, healthcare institutions and application developers can streamline processes such as automated report generation, extraction of insights from unstructured records, and development of intelligent medical solutions—eliminating the need to build specialized AI models independently. Their domain-specific training ensures superior accuracy and reliability in clinical contexts compared to generalized AI systems.
VibeFlow is an advanced AI-driven full-stack application generator designed to convert natural language prompts into fully functional web applications, complete with both frontend and backend components. Unlike conventional no-code platforms, VibeFlow offers complete code transparency and full editability, enabling users to review, understand, and modify every aspect of the generated application. By leveraging artificial intelligence, the platform accelerates application development without compromising developer control. Key features include a visual backend editor—similar to UI design tools like Figma—and built-in database and service integrations. Unlike tools that produce only static layouts, VibeFlow delivers applications with real business logic, providing significant value for non-technical users seeking functional and scalable solutions.
Drivetrain is an advanced AI-powered Autonomous FP&A platform designed to transform financial planning and analysis for mid-market and enterprise organizations. The platform automates the creation of financial models, budgets, and forecasts by seamlessly extracting and processing data from ERP, CRM, and HRIS systems, eliminating the need for manual spreadsheet operations. What differentiates Drivetrain is its deep AI integration, featuring natural language-driven data transformation, a conversational analytics interface for complex financial queries, and continuous AI monitoring that detects anomalies in real time. Finance leaders adopt Drivetrain to reduce spreadsheet errors and consolidate data from more than 200 business tools into a centralized platform.
SmythOS is an advanced no-code AI platform designed to enable users to build, deploy, and manage intelligent AI agents without writing any code. Through its AI-powered architecture, the platform empowers users to create custom automation workflows and virtual agents using an intuitive drag-and-drop interface. This visual builder allows the design of complex AI tasks, including customer support, content generation, and data processing, with minimal effort. SmythOS differentiates itself through seamless integration with leading AI technologies such as OpenAI and Hugging Face, and its natural language instruction capabilities that let users specify functionality through simple descriptions. Businesses and individuals leverage SmythOS to accelerate AI adoption, automate repetitive operations, and deploy sophisticated AI-driven solutions across diverse environments.
