Weekly Piece of Future #173
From Ultrasound Wristbands to Physics-Aware AI and Quantum Self-Testing Chips
Hey there, fellow future-addicts!
Welcome to this week's edition of Rushing Robotics, where the boundary between science fiction and reality dissolves a little more each week. As we cross into mid-June 2026, the pace of innovation feels less like progress and more like propulsion—each breakthrough building on the last in ways that redefine what we thought possible. From robots that read our muscles to chips that verify their own quantum integrity, this week's discoveries remind us that the future isn't arriving gradually. It's arriving in waves.
Unlock truly uncensored AI—privately.
While most AI platforms store and analyze your conversations on their servers, Venice keeps everything local to your browser—meaning your private thoughts and queries stay truly private with no server-side storage or data harvesting.
Venice leverages Trusted Execution Environments (TEEs) for end-to-end encrypted processing—ensuring that even during computation, your data remains cryptographically protected from prying eyes.
Get a $10 credit when you register through my link. Upgrade to Venice Pro and apply the credit toward the most advanced models immediately. No filters, no logs—just pure creative freedom.
Claim your $10 credit & start creating!
[Disclosure: I earn a commission if you sign up via this link.]
🤯 Mind-Blowing
This section features the kind of breakthroughs that make you reconsider what's possible. From wearable ultrasound systems that give robots human-like dexterity, to AI that learns physics before training, to quantum chips that verify their own integrity—these are the stories that push the boundaries of imagination. We also explore Hong Kong's first humanoid-run convenience store and a photomemristor that helps self-driving cars see through blinding light. The future isn't waiting for permission.
🔊 Industry Insights & Updates
The intersection of big tech and real-world deployment takes center stage here. NVIDIA and LG are weaving AI into factory floors and household robots, while French startup C12 advances quantum computing with carbon nanotube assembly. MIT's diamond-embedded chips could unlock 6G communications, and GigaAI has placed 100 humanoid robots into actual Chinese households for the first large-scale real-home test. The gap between lab and living room is closing fast.
🧬 BioTech
Life sciences are experiencing their own revolution. An AI model called TITO can now predict molecular evolution over 10,000 times faster than conventional methods, potentially transforming drug discovery. Zinc oxide nanowires are enabling liquid biopsies that capture cancer signals with 90% efficiency. And engineered extracellular vesicles are delivering full-length mRNA to restore dystrophin production in Duchenne muscular dystrophy—offering hope where viral vectors fell short. The code of life is becoming increasingly writable.
💡 Products/Tools of the Week
For builders and makers, this week brings some powerful new tools. Nous Research released Hermes Desktop, bringing the open-source Hermes Agent to macOS, Windows, and Linux with persistent memory, multimodal AI tools, and sandboxed execution. Archyl is redefining software architecture with AI-powered "architecture as code" that auto-generates interactive C4 diagrams and detects drift. TukiAI launched an AI sales and service agent for WooCommerce that syncs inventory in real-time and supports 54 languages across web, WhatsApp, and social channels. And Kuberns debuted an AI-powered cloud PaaS that automates end-to-end deployments, handling everything from CI/CD pipelines to cost optimization—no DevOps expertise required.
🎥 Video Section
See the future in motion: Google's Gemma Playground brings us a robot duck, UBTECH's U1 Pro humanoid robot blurs the line between human and machine, and Unitree's robots leave Simon Cowell speechless on America's Got Talent 2026. Sometimes you just have to see it to believe it.
There's something deeply thrilling about watching so many disparate threads—quantum hardware, embodied AI, biotech delivery systems, physics-aware neural networks—converge toward a future that feels both imminent and extraordinary. We're not just building smarter tools; we're building systems that understand the world's fundamental rules and operate within them more elegantly than we ever imagined. The question is no longer whether these breakthroughs will reshape our lives, but how quickly we'll adapt to a world where the impossible becomes routine. Stay hungry, stay futurish!
🤯 Mind-Blowing
A new wearable ultrasound wristband has been developed to give humanoid robots human-like dexterity by capturing detailed muscle and tendon activity beneath the skin. Researchers at MIT, led by Xuanhe Zhao, created the device to address the limitations of camera-based systems and strain sensors that struggle with continuous finger motions. The wristband uses a 256-channel wireless ultrasound imaging system and a hybrid Transformer-ResNet AI model to continuously track all 22 degrees of freedom of the human hand. In laboratory tests with eight volunteers, the system reproduced hand gestures with high accuracy and low latency, mirroring movements within 120 milliseconds and successfully recognizing all 26 letters of American Sign Language. The wireless technology allows users to remotely control robotic systems, which the team demonstrated by directing a robotic hand to play the piano.
A machine learning system that learns the laws of physics before training has been developed to design advanced optical materials up to ten times faster than conventional methods. Researchers at Chalmers University of Technology embedded the fundamental laws of electromagnetism directly into the neural network, allowing the system to understand how light and electromagnetic fields behave from the start. This approach avoids forcing the model to discover these principles from scratch and significantly reduces the time required for data generation. According to Philippe Tassin, a professor at the Department of Physics and Astronomy, feeding the super-brain information about physics made it much smarter, reducing calculation times to one tenth of what was previously required.
Hong Kong's first humanoid robot-run convenience store is set to open on the Hung Hom waterfront. A mainland Chinese firm specializing in embodied AI developed the 24-hour outlet, which features a robot nicknamed Xiao Gai that can interact directly with customers, initiate friendly conversations, and support purchases in multiple languages. The nine-square-meter capsule format can be configured to sell snacks, creative merchandise, and over-the-counter medicines. Financial Secretary Paul Chan revealed the project in his weekly blog, framing it as an endorsement of the city's open market and its role as a showcase for global innovation.
A self-testing quantum random number generator chip has been developed to verify the integrity of its own hardware while producing random numbers. Researchers at the National University of Singapore, led by Associate Professor Charles Lim, created the device to eliminate the need to trust that measurement components are operating as specified. Using a measurement-device-independent protocol, the chip generates known quantum light states and compares detector responses against quantum theory predictions, stopping automatically if results do not match. The device integrates both the signal encoder and optical detector on a single silicon chip that operates at room temperature, achieving a detector efficiency of 69.1 percent and producing 64 certified random bits per second.
A new photomemristor device mimicking the human eye has been developed to prevent self-driving cars from going blind in high-contrast lighting. Penn State researchers created the component to handle the transition from blinding brightness to deep shadow almost instantly, unlike the human eye which takes 20 to 30 minutes to acclimate. The device works by pairing titanium oxide with a stretchy plastic that swells in the dark and dries in bright light, acting as a physical volume knob for light sensitivity. This innovation allows autonomous vehicles to distinguish details like the glow of a red light against bright headlights, solving a major fragility in current robotic vision systems.
🔊 Industry Insights & Updates
A connected AI ecosystem linking model development, simulation, and deployment has been unveiled through a broad collaboration between NVIDIA and South Korea's LG Group. The partnership aims to weave AI into factory floors, household robots, and autonomous vehicles by combining NVIDIA's accelerated computing platforms with LG's manufacturing and consumer electronics expertise. LG will use NVIDIA's digital twin technologies to create an autonomous manufacturing ecosystem that connects procurement, production, and logistics through real-time decision-making. For robotics, LG plans to incorporate NVIDIA Isaac Sim, Isaac Lab, and the GR00T foundation model to train home robots in virtual environments before deployment. The collaboration also extends to infrastructure, with LG Electronics working on liquid-cooling technologies for the NVIDIA DSX AI factory platform and LG Uplus building large-scale AI data centers.
A nanoassembly process for transferring individual carbon nanotubes onto quantum chips has been advanced by French quantum computing startup C12 through a technique called pick-and-place. The company's Quantum Processing Unit integrates carbon nanotubes made from ultra-pure carbon-12 isotope on semiconductor chips, where these nanotubes house spin qubits. This development allows C12 to preselect and qualify individual carbon nanotubes before integration, enabling tighter quality control of assembled devices.
A new chipmaking technique that embeds gallium nitride transistors into an ultrathin diamond layer has been developed by MIT researchers to improve high-power wireless systems. This method allows the diamond to act as a heat spreader within the chip, normalizing the temperature across the device so transistors can approach peak performance without reliability issues. The team built a power amplifier using this approach that achieved higher output power, efficiency, and gain than similar devices in existing literature. This advancement addresses the thermal management bottleneck that has previously limited the use of gallium nitride in demanding applications like 6G and satellite communications.
A large-scale real-home test for general-purpose humanoid robots has been launched in China by Wuhan-based GigaAI, which deployed the first batch of 100 SeeLight S1 units into actual households. The robots successfully performed chores like preparing breakfast, loading dishwashers, and folding laundry after less than a month of on-site training. While the company utilizes an embodied foundation model that allows the machines to process natural language and adapt to changing environments, current limitations include slow execution speeds and difficulty handling delicate items like cups without spilling. GigaAI views this deployment primarily as a data-collection platform to improve future models.
🧬 BioTech
A new artificial intelligence model capable of predicting molecular evolution more than 10,000 times faster than conventional simulations has been developed to accelerate drug discovery. Researchers from Chalmers University of Technology and the University of Gothenburg created the TITO model to fast-forward through molecular dynamics, bypassing the need to calculate forces step by step over billions of femtoseconds. By learning the underlying rules of molecular motion from over 12,500 organic molecules and short peptides, the model can predict how new molecules will behave and change over time scales a thousand times longer than its training data. This breakthrough allows researchers to identify promising drug candidates more quickly and accurately without relying on computationally demanding numerical calculations.
A new cancer detection device using zinc oxide nanowires has been developed to selectively capture extracellular vesicles from bodily fluids for liquid biopsies. The group, led by Nagoya University professor Takao Yasui and collaborating with Hokkaido University professor Yasuhide Inokuma, solved the challenge of attaching antibodies to nanowires by using a synthetic polymer variant called pKNHS 4.2. This specific polymer allows for single-step antibody modification, which enabled the device to achieve 90 percent capture efficiency for target molecules in cultured breast cancer cells. When tested on serum from ovarian cancer patients, the antibody-conjugated nanowires successfully isolated cancer-related vesicles and revealed distinct microRNA profiles, keeping the molecular contents intact for sensitive disease analysis.
A new treatment platform using engineered extracellular vesicles has successfully delivered full-length messenger RNA to restore dystrophin production in preclinical models of Duchenne muscular dystrophy. Researchers at The University of Texas MD Anderson Cancer Center developed the natural nanoscale delivery particles to overcome the size limitations of current viral-based gene therapies, which cannot carry the entire DMD gene. The engineered vesicles were tagged to target skeletal muscles directly after bloodstream injection, resulting in improved muscle strength and function without the serious side effects or immune reactions typically associated with viral vectors. This breakthrough suggests that extracellular vesicle-mediated mRNA therapeutics could serve as a blueprint for restoring large proteins in other degenerative diseases.
💡Products/tools of the week
Nous Research released Hermes Desktop, bringing the open-source Hermes Agent to macOS, Windows, and Linux with a polished chat interface. The application shares persistent memory, sessions, API keys, and skills with the CLI and web gateway, ensuring work context syncs across all interfaces. Users can invoke multimodal AI tools like web search and image generation, schedule recurring automations, and spawn isolated subagents for delegated tasks. It supports various backends including local, Docker, and Modal for hardened, sandboxed execution.
Architecture as code is now possible through a new AI-powered platform. Archyl released a living software architecture documentation tool that uses AI to analyze codebases and automatically discover components. It generates interactive C4 diagrams and detects drift so documentation remains accurate. By expressing architecture as code via a YAML DSL, Archyl enables integration with Git workflows for versioning, reviews, and collaboration. Architects, tech leads, and dev teams can now replace manual, stale diagrams with faster, more accurate, and reviewable architecture docs.
Real-time inventory synchronization and multilingual customer support are now available for WooCommerce stores. TukiAI released an AI sales and service agent as a WordPress plugin that leverages leading LLMs such as GPT-4, Claude, Gemini, and Mistral to read catalogs, sync prices, and manage orders. The platform operates 24/7 across websites, WhatsApp, and social channels, answering questions in 54 languages and offering context-aware product recommendations. TukiAI programmed the agent to escalate to humans when unsure, helping store owners automate repetitive support, recover abandoned carts, boost conversions, and free their teams for higher-value tasks.
Kuberns launched an AI-powered cloud PaaS that uses agentic AI to automate end-to-end deployments and operations. The platform auto-detects applications, generates configs, and wires CI/CD pipelines automatically. Kuberns handles scaling, monitoring, secrets, and cost optimizations so teams can ship in minutes without needing DevOps expertise. Users leverage the platform to cut deployment time, reduce cloud spend, and access production-grade automation like live logs and alerts with minimal setup.
