Weekly Piece of Future #127
From Machine Taste to AI Browser and Surgical Robots
Hey there, fellow future-addicts!
Welcome to this week's edition of Rushing Robotics! The pace of innovation is accelerating. This week, life-changing medical breakthroughs, smarter AI, and next-gen robotics showcase how the future is becoming reality right now. From restoring movement to the paralyzed, to AI that tastes and robots that assist surgeries — the boundaries of what’s possible keep shifting.
🤯 Mind-Blowing
Imagine a child who lost the ability to walk—then walks again thanks to a new treatment that revives cellular energy. A rare mitochondrial disorder met its match. Meanwhile, brain implants are evolving: one company’s less invasive approach already outnumbers Neuralink’s implants. Machines are learning to taste with near-human precision using graphene sensors. Lab-grown organs now grow blood vessels for the first time, bringing us closer to transplant-ready organs. And humanoid robots have taken their first real steps in performing medical tasks.
🔊 Industry Insights & Updates
The U.S. just launched Nexus, a 400-petaflop AI supercomputer built to supercharge science and research. Meta is investing billions to build colossal AI superclusters, aiming for unprecedented computing power. In manufacturing, Apera’s Forge platform enables rapid AI training of vision-guided robots — all in a web browser. And a new SSD with a built-in self-destruct system promises ultimate data security for sensitive industries.
🧬 BioTech
Memory might be stored like a tree—compressed summaries with detailed branches—offering new insights into how we recall stories and events. Surgeons in Australia rehearsed and performed a complex aortic replacement with help from a custom 3D-printed model of a patient’s artery. The UK Biobank completed whole-body scans on 100,000 volunteers, creating one of the richest datasets ever to understand disease progression and human health.
💡 Products/Tools of the Week
Say hello to Comet, an AI-powered browser assistant that makes researching and multitasking online effortless by summarizing pages and handling complex commands. For developers, Forge Code offers AI pair programming right inside your terminal—boosting productivity while keeping your code private and secure. Content creators will love Perso, a smart dubbing platform that flawlessly translates and lip-syncs videos into over 30 languages, cutting localization time and costs dramatically. And for 3D artists, Rendable3D uses AI to turn simple inputs into detailed, Blender-ready models, speeding up creative workflows without sacrificing control.
🎥 Video Section
Watch humanoid robot MagicBot Z1 in action, explore 1X Robotics’ Palo Alto HQ, see quadrupedal robots tackle tough terrain, meet household humanoids from NEURA Robotics, and check out Kinisi’s autonomous warehouse robots.
The breakthroughs you just read about are just the beginning of what’s possible when human ingenuity meets technology. There’s no telling how far these innovations will take us, but one thing is clear: the future belongs to those who dare to imagine. Stay hungry, stay futurish!
🤯 Mind-Blowing
An eight-year-old boy diagnosed with HPDL deficiency—a rare and fatal mitochondrial disorder—has regained the ability to walk following a novel treatment that revives cellular energy production. HPDL deficiency disrupts the biosynthesis of coenzyme Q10 (CoQ10), an essential molecule for mitochondrial energy metabolism. The resulting energy failure leads to rapid-onset paralysis, muscle stiffness, and profound fatigue. Within three months of diagnosis, the boy had lost the ability to walk and was confined to a wheelchair. In a pioneering clinical effort, researchers at NYU Langone Health administered an experimental compound designed to restore CoQ10 production. The treatment not only halted the progression of the disease but also reversed the child’s paralysis—marking the first real-world application of a biochemical approach to targeting mitochondrial dysfunction in HPDL deficiency. This breakthrough may pave the way for future therapies for similar mitochondrial diseases.
A lesser-known rival to Neuralink is making quiet but significant progress in brain-computer interface technology. Synchron has implanted its Stentrode device in 10 patients since 2019, slightly ahead of Neuralink’s seven. The key difference lies in the surgical method. The Stentrode doesn’t require opening the skull—it’s inserted into a blood vessel above the brain’s motor cortex, avoiding direct contact with brain tissue. This design greatly reduces surgical risks but introduces a major challenge: the signal strength is inherently weaker due to the increased distance from neurons. Synchron, however, is undeterred. COO Kurt Haggstrom dismissed Neuralink’s complexity, saying, “We think Neuralink is over-engineering. I have no doubt this will work.” The device is expected to cost around $40,000 to $50,000 and may represent a more scalable path to BCI adoption for patients who need assistance with mobility or communication.
Scientists have developed a graphene-based sensor capable of tasting with near-human precision, marking a significant advancement in artificial sensing technologies. The device employs machine learning to analyze chemical signals and classify flavors, including those not previously encountered. A key innovation lies in its functionality under moist conditions—a first for artificial gustatory systems—allowing the sensor to more accurately mimic the environment of human taste buds. Constructed from layered graphene oxide in a nanofluidic structure, the device integrates both sensing and computation on a single platform, improving efficiency over earlier artificial tongue designs. Graphene oxide, like its pure form, exhibits electrical responses when exposed to various chemicals. The researchers trained the system using signals from 160 compounds representative of common flavor profiles. These inputs enabled a machine-learning algorithm to learn how each flavor affects the material’s conductivity, forming a robust memory for precise flavor detection.
Two independent research teams have made a major breakthrough in regenerative medicine: successfully growing blood vessels within lab-generated organs. This achievement addresses one of the most persistent challenges in creating functional organoids—tiny, lab-grown replicas of organs such as hearts, livers, and lungs. While organoids have become vital tools for studying diseases and testing new drugs, their development has been limited by the absence of vascular networks. Without blood vessels, these miniature organs cannot receive the oxygen and nutrients needed for proper growth, function, or maturation. For example, mini-kidneys have been unable to filter blood, and mini-lungs cannot exchange gases effectively in the absence of vascular systems. A team at the University of California initially tried building the organ and vessels separately. This method led to organoids with better 3D structure, enhanced cell diversity, improved survival rates, and more advanced tissue development than previously seen. The innovation marks a significant step toward creating fully functional, transplant-ready organs in the future.
As hospitals grapple with overcrowded waiting rooms, physician burnout, and increasing surgical backlogs, humanoid surgical robots may offer a scalable solution. Researchers at the University of California, San Diego, contend that current surgical robots are highly specialized, expensive, and reliant on expert operators, which restricts their broader clinical deployment. To address this, the UC San Diego team developed a bimanual teleoperation system for the Unitree G1 humanoid robot and evaluated its performance across seven diverse medical procedures, ranging from physical exams to emergency interventions and needle-based precision tasks. Results demonstrated that humanoid robots are capable of replicating critical components of clinical workflows, including tasks like ventilation and ultrasound-guided procedures. However, the study also identified limitations, particularly in executing high-strength maneuvers and maintaining sensor accuracy, indicating that further refinement is needed for widespread adoption.
🔊 Industry Insights & Updates
The U.S. has unveiled Nexus, a cutting-edge AI supercomputer capable of processing data faster than 8 billion humans combined. Built to support the country’s research community, Nexus delivers an extraordinary 400 quadrillion operations per second (400 petaflops), along with 330 terabytes of memory and 10 petabytes of flash storage. Designed specifically for artificial intelligence and high-performance computing (HPC) workloads, Nexus will accelerate advancements in fields such as drug discovery, clean energy, climate modeling, and robotics. Beyond raw performance, Nexus also emphasizes accessibility. Its user-friendly interfaces are engineered to make advanced AI tools easier to use for scientists and researchers across a wide range of disciplines, bridging the gap between technical computing and real-world problem-solving.
Meta Platforms Inc., the parent company of Facebook, is significantly ramping up its artificial intelligence infrastructure efforts. CEO Mark Zuckerberg announced plans to invest “hundreds of billions of dollars” into constructing some of the largest AI superclusters the world has ever seen. In a statement shared via Facebook and Threads, Zuckerberg revealed that the company’s first AI-focused data center, named Prometheus, is scheduled to go online in 2026. He also mentioned a second facility, Hyperion, which will have the capacity to scale to 5 gigawatts over time. According to industry analysis from SemiAnalysis, Meta is on track to become the first AI research organization to deploy a supercluster with more than one gigawatt of computing power.
Apera Forge, developed by the Canadian firm Apera, is a web-based AI vision and simulation platform designed to accelerate the deployment of 4D vision-guided robotic systems. Operating entirely in the browser, Forge eliminates the need for hardware during the design and training phases, allowing for rapid configuration of robotic cells. Key features include support for EOAT-mounted cameras and de-racking scenarios, enabling manufacturers to simulate full automation environments—including robot arms, grippers, parts, and workspace configurations—in a matter of minutes. The system then runs over a million virtual training cycles to develop a robust AI neural network capable of delivering >99.9% task reliability. By compressing development timelines from weeks or months to mere hours, Apera Forge enhances operational agility and reduces implementation risks—making it a strategic asset for manufacturers pursuing advanced VGR (Vision-Guided Robotics) capabilities at scale.
Taiwanese storage manufacturer TeamGroup has introduced the P250Q-M80, a next-generation internal SSD with a built-in self-destruct mechanism, designed to offer unparalleled data security for industries where confidentiality is critical. The device targets sectors such as defense, industrial automation, AI development, and high-stakes crypto storage. At the core of the P250Q-M80 is a dual-mode self-destruction system. A short press of the external control initiates a software-level data wipe, preserving the drive hardware while erasing all user data. This feature includes an auto-resume mechanism, which continues the deletion process if power is interrupted mid-operation. For more extreme scenarios, the SSD features a hardware-level kill switch. Holding the external red button for 5 to 10 seconds sends a high-voltage surge into the NAND flash chips, physically destroying the memory. A multi-stage LED status indicator gives users visual feedback, confirming each stage of the self-destruction sequence.
🧬 BioTech
A team of scientists has developed a novel model to explain how the human brain stores narratives, proposing that memory is organized as a tree-like structure. In this model, nodes near the root encode compressed summaries, while branches represent increasingly detailed information. The approach relies on random trees—mathematical constructs used to model branching structures—to simulate how people encode and recall stories. According to the study’s lead authors, the goal was to establish a formal, mathematically grounded theory of memory for complex narrative material. Drawing from mathematics, computer science, and physics, their framework examines how events and narrative elements are interconnected in memory. The researchers suggest their findings may offer broader insights into human cognition, particularly given the central role that narratives play in how individuals understand personal experiences, as well as social and historical phenomena.
Surgeons at Brisbane’s Prince Charles Hospital have successfully performed one of Australia’s most complex vascular surgeries, replacing nearly the entire aorta of a man in his late 50s using a synthetic graft. The procedure was made possible through the use of a custom, life-sized, multi-material 3D model of the patient’s enlarged aorta, allowing the surgical team to rehearse and plan the intricate operation in advance. Scans had revealed that the man's aorta—the body’s largest artery—had expanded to approximately four times its normal diameter, posing an imminent risk of rupture. Engineers and industrial designers at the Herston Biofabrication Institute created the detailed anatomical model using advanced computer imaging and patient-specific data, enabling precise surgical preparation. The successful outcome highlights the growing role of biofabrication technologies in enhancing patient outcomes in high-risk procedures.
The UK Biobank has successfully concluded the world’s largest whole-body imaging study, capturing detailed scans of the brains, hearts, abdomens, blood vessels, bones, and joints of 100,000 volunteers. This record-breaking effort offers scientists an unparalleled look into the human body across the aging process, providing critical insight into when, why, and how diseases develop. Since 2015, UK Biobank has released imaging data in stages, enabling researchers worldwide to make major advances in diagnostics for conditions such as heart disease, cancer, and dementia. Soon, qualified researchers will gain access to over one billion anonymized images. When combined with 15 years of extensive data on participants’ lifestyle, medical history, genetics, and blood biomarkers, these scans create an extraordinarily rich resource. For the first time, scientists can observe in detail how diverse biological and environmental factors influence health over time, gaining new clarity on rare conditions and tracking the progression of common diseases. The project also enhances researchers’ ability to compare healthy individuals to those with complex health issues—paving the way for stronger, earlier indicators of disease.
💡Products/tools of the week
Comet is an AI-centric web browser created by Perplexity, engineered to redefine the browsing experience through comprehensive integration of artificial intelligence. Departing from traditional browsers, Comet introduces a robust AI assistant embedded in the sidebar that can summarize web pages, respond to user inquiries about on-screen content, and carry out complex operations across multiple websites. Users can interact with Comet using natural language commands—whether typed or spoken—to manage tabs, perform research, and automate online processes. Its AI capabilities enable agentic actions such as comparing products, booking travel, or synthesizing research without requiring manual tab-switching.
Forge Code: The AI Shell is a terminal-native AI pair programmer designed to let developers interact with their codebase using natural language—all without leaving their preferred development environment. Seamlessly integrated into existing workflows, this tool delivers intelligent support for tasks such as debugging, code generation, refactoring, code review, and Git operations via conversational commands. What truly distinguishes Forge Code is its emphasis on privacy and security: all code and AI-driven analysis are processed locally. Additionally, developers can tailor performance with a range of AI models, optimizing for speed, context window size, or cost depending on project needs. Whether used by solo developers or large engineering teams, Forge Code boosts productivity, streamlines repetitive processes, and simplifies navigation through complex codebases.
Perso is an AI-driven dubbing and localization platform that enables seamless translation and lip-syncing of video content into more than 32 languages with exceptional accuracy. Built for content creators, educators, and organizations, Perso automates critical processes such as multi-speaker recognition, voice cloning, and realistic lip synchronization to deliver high-quality dubbed videos. The platform also allows users to modify translated scripts in real time, offering the flexibility to refine tone, correct language nuances, and ensure terminological accuracy. Supporting a wide range of formats and durations, Perso is particularly well-suited for content such as interviews, podcasts, talk shows, and short videos. By significantly reducing localization costs and timelines, Perso provides a scalable solution for professional-grade video adaptation—without requiring traditional voiceover or filming infrastructure.
Rendable3D is an advanced AI-powered platform designed to generate high-quality 3D models compatible with Blender, eliminating the need for extensive modeling expertise. By utilizing artificial intelligence, the tool converts basic inputs into detailed, production-ready 3D assets, accelerating traditional modeling workflows. It enables designers, animators, and game developers to bypass labor-intensive manual processes while maintaining creative control through seamless integration with Blender. By automating foundational modeling tasks, Rendable3D empowers users to shift their focus from technical complexity to artistic refinement, offering a faster, more accessible route to professional-grade 3D content creation.
MagicLab: MagicLab Presents MagicBot Z1: Next-Gen Bipedal Humanoid Robot
Arc Lab: Learning stable bipedal locomotion skills for quadrupedal robots on challenging terrains
NEURA Robotics: Discover 4NE-1: The era of household humanoids is coming soon.
Kinisi Robotics: The Future of Warehousing Starts Now: Meet Kinisi’s Autonomous Humanoid Robot