Sodium-ion batteries have a compelling pitch: sodium is cheap and abundant, the chemistry has attractive safety characteristics, and it sidesteps much of the supply-chain anxiety around lithium. Their weakness has been energy density. A new study from Xi'an Jiaotong University, published in Nature Energy on November 2, 2026, tackles that problem with a radical design choice: remove the anode entirely, and then fix the chemistry that has kept such batteries from lasting.
According to the univers...
More
Graphene Harvester Runs Sensors on Power-Line Energy
Every wall in a modern building hides a quiet, constant source of energy that nobody uses. The copper wiring that carries household and office electricity produces a weak, oscillating magnetic field around it, sixty times a second in North America. Researchers at the University of Arkansas and the University of Michigan have now shown that this stray field can do useful work: it can keep a tiny temperature sensor running with no battery at all.
According to the University of Arkansas announceme...
More
Empatica’s WellPD Takes Parkinson’s Monitoring Home
Parkinson's disease does not hold still for an appointment. Symptoms shift hour by hour as medication takes hold and wears off, which means a neurologist's twenty minutes with a patient captures one arbitrary slice of a moving picture, and the rest of it has to be reconstructed from memory. On October 28, the Boston- and Milan-based wearables company Empatica launched a product aimed squarely at that gap: WellPD, an FDA-cleared and CE-marked wrist-worn device and companion app that tracks Parki...
More
UCSF’s B-SYNC CAR-T Cells Switch On Only in the Brain
Diffuse midline glioma is the kind of diagnosis that ends conversations. It grows into the brainstem, the piece of tissue that runs breathing, heartbeat and swallowing, and it is one of the few brain cancers that mostly strikes children. Surgeons cannot cut it out. Most children die within a year. Engineered immune cells, the technology that transformed blood cancers over the past decade, have repeatedly failed to dislodge it.
On October 28, UC San Francisco announced a result that may change t...
More
Samsung Galaxy Buds On: Open-Ear Earbuds Arrive in the US
For years, wireless earbuds have competed on one thing above all: shutting the world out. Bigger seals, stronger noise cancellation, deeper isolation. But a growing group of listeners wants the opposite, at least some of the time. Runners need to hear traffic, office workers want to catch a colleague's question, and commuters want a podcast without missing a station announcement. On October 28, 2026, Samsung announced U.S. availability of Galaxy Buds On, its first open-ear earbuds, built for ex...
More
Pudu TOMO S: The Robot Mower That Sharpens Its Own Blades
Robot lawn mowers have quietly become one of the most successful home robots after the vacuum. They map the garden, trundle out on schedule and dock themselves when the job is done. Yet one chore has stubbornly stayed with the homeowner: dealing with the blades. Dull blades tear grass rather than slice it, and swapping or sharpening them means handling sharp metal under a machine. On November 3, 2026, Pudu Robotics, a company better known for the delivery and cleaning robots used in restaurants...
More
Flower’s Endeavor 1.1 Puts Frontier AI Under Customer Control
Companies adopting AI keep running into the same trade-off. The most capable models live behind closed APIs, where the provider decides what changes, where the computing happens and what customers may do with the results. Running an open model yourself brings control, but it also brings the burden of hosting it and understanding how it behaves. On November 3, Flower Labs announced a model meant to remove that choice. According to the company's announcement, Endeavor 1.1 is a frontier model that...
More
NVIDIA RTX Spark Brings On-Device AI Characters to Games
For most of gaming history, the people you met inside a game could say only what a writer had scripted for them. Ask a guard something off-script and you got silence or a canned line. NVIDIA wants to change that, and its latest developer update shows how close the company thinks we are. In an October 27 post on the NVIDIA Technical Blog, the company laid out a bundle of AI tools for its new RTX Spark PC platform: a small language model built to run game characters on the device itself, a faster...
More
KIT Prints a Perovskite-Silicon Tandem Cell Past 30% Efficiency
Printing a 30% solar cell
Perovskite-silicon tandem solar cells are the leading candidate to push panels past the limits of silicon alone, but many record cells are made with lab-scale methods that do not translate easily to factories. According to TaiyangNews, researchers at Germany's Karlsruhe Institute of Technology (KIT) have now demonstrated a perovskite-silicon tandem cell with 30.2% efficiency using a process built around inkjet printing, which they describe as the first time such a cell...
More
Organic Solar Cell Splits Water at a Record 7.9% Efficiency
A solar cell that makes hydrogen without a power supply
Hydrogen made from water and sunlight is one of the cleanest fuels on paper, but most designs still lean on an outside electrical boost. According to a report by Interesting Engineering, a team at South Korea's Ulsan National Institute of Science and Technology (UNIST) has built an organic-semiconductor device that splits water entirely on its own, and set an efficiency record for its class while doing it.
The headline number is a 7.9% sol...
More
NVIDIA’s 9.6M-Parameter RE-USE Cleans Speech in Any Language
Most speech-cleanup models are big. NVIDIA's new NemotronLabs RE-USE is not: it has 9.6 million parameters, yet its model card shows it matching or beating open models that are 26 to 50 times larger on several measures. NVIDIA researchers promoted the open model on October 30, 2026, according to GIGAZINE.
What it does
According to NVIDIA's Hugging Face model card, RE-USE is a multilingual universal speech enhancement model. It restores speech damaged by noise, reverberation, clipping, bandwidth...
More
KAIST’s CluSTER Cuts AI Fine-Tuning Time by 70%
Training a specialist AI model is expensive mostly because of repetition. Researchers at KAIST say they have found a way to cut that repetition: a method called CluSTER fine-tuned a model on medical question-and-answer data using about 31 percent of the dataset and cut training time by 69.6 percent, with only a small loss in accuracy, Aju Press reported on November 2, 2026.
The problem: fine-tuning wastes GPU time
Large language models first learn broad knowledge, then developers fine-tune them...
More
Nano-Wrinkles and a Tiny Actuator Fight Implant Biofilms
Infections on implantsJoint replacements, pacemakers and catheters are prone to biofilms, bacterial sheets that can cause hard-to-treat, antibiotic-resistant infections. A University of Michigan Engineering team built an implantable platform, DARTS (dynamically actuated reconfigurable topographical surface), that reduced bacterial burden and improved tissue response in mice, according to the university's October 27 news release. The study appeared in Advanced Functional Materials.Two mechanisms...
More
A Mesh Grown Into Heart Cells Turns Heartbeats Into Power
Power from the body itselfImplants run on batteries that eventually die. Engineers at the University of Massachusetts Amherst grew living human heart tissue around a mesh of tiny generators, and each heartbeat bent the generators enough to produce an electrical pulse, Earth.com reported on October 23. The team's calculation puts the output at more than ten times the energy for the device's size compared with earlier implanted harvesters. None of it has been tried in a body.DesignThe mesh is abo...
More
MIT’s LUCID Implant Uses Ultrasound to Spread Drugs in the Brain
The problemGlioblastoma is an aggressive brain tumor; MIT News notes a little more than 10,000 cases are diagnosed in the US each year and patients survive 12 to 15 months on average. Surgeons struggle to remove every cancer cell, so tumors often recur. Implantable wafers and gels can release drugs locally, but they rely on passive diffusion.What MIT builtResearchers led by Canan Dagdeviren created LUCID, the Long-term Ultrasound-enhanced Convective Injection Device, described in a paper in the...
More
A 3D-Printed Brain Implant Cools Seizure Activity in Rats
One printed device, three jobsResearchers at the University of Sheffield and TU Dresden have printed a brain implant that records electrical activity and cools a patch of cortex, all as one continuous structure, according to a Tech Times report on October 29. The work was published in Advanced Science in September 2026. The aim is an alternative approach for drug-resistant focal epilepsy, which affects roughly 15 million people worldwide per the report.How it is builtUsing direct ink writing, a...
More
Penn State’s Printable Hydrogel Targets Multitasking Wearables
A gel that sticks where others slipMost wearable health monitors capture one signal at a time, and they struggle with hair, sweat and movement. A team led by Penn State has developed an ultrasoft, adhesive, printable hydrogel designed to conform to dry, wet, moving and hairy skin, Medical Product Outsourcing reported on October 30. The work was published in Science Advances.What is in itThe researchers call it RTLR gel, a name that refers to its radically tuned setting time and two forms of gra...
More
NTNU’s Color-Changing Patch Reads Pulse Without Electronics
A rainbow patch that reads your heartbeatResearchers at the Norwegian University of Science and Technology (NTNU) have built a small transparent patch that looks like the back of a CD and, paired with an ordinary smartphone camera, can measure pulse and blood vessel stiffness. Medical Design & Development, which republished the university's story on October 30, calls it the first of its kind. The patch is silicone rubber with no electronics or cables.How it worksOne side carries a nanoscale pat...
More
Mischief 1: A Video Model Tuned by Lab-Grown Neurons
Brain cells in the lab, software on the GPUOn October 30, The Biological Computing Co. (TBC) launched mischief 1, an image-to-video model it says generates video five times faster than its base model. It is available through fal and OpenRouter and starts at half a cent per second, so a five-second clip costs 2.5 cents and 100 clips cost $2.50, according to the company's release.The unusual part is the origin story. TBC grows living neurons in its lab and studies how they respond to signals. It ...
More
Mistral Large 4 Goes Open-Weight at 1 Trillion Parameters
What was releasedOn October 29, Mistral AI released the weights of Mistral Large 4, nicknamed Le Chonk, three weeks after a public preview. According to Hugging Face's launch post, it is a multimodal, multilingual mixture-of-experts model with about 1 trillion total parameters, about 49 billion active per token, and a 1 million token context window. It takes text and images in and produces text.Two official variants are available, FP8 and NVFP4, plus community GGUF builds. The Next Web reports ...
More
One Robots.txt Rule to Opt Out of AI Training
A single switch instead of manyOpenAI, Google, Microsoft and Anthropic said on October 30 that they will adopt a new standard for publishers who do not want their pages used to train AI, The Verge reported. Until now, a site had to signal its wishes to each company separately. The new control lets a publisher apply one preference across multiple AI companies.The proposal comes from Kevin Kelley, a technical lead for standards at Anthropic, and moves through the Internet Engineering Task Force, ...
More
DeepSeek’s Super Agent Aims to Manage Itself
A team of AIs behind one chat windowOn October 31, DeepSeek published a new architecture it calls Super Agent, according to a November 1 report by The Asia Business Daily. The pitch is simple: today's AI assistants answer one question at a time, but the people supervising them can only follow so much. DeepSeek says a person can manage roughly three to five conversations at once, which caps how much work AI can actually absorb.Super Agent tries to lift that cap by turning a single assistant into...
More
China Zorrilla: World’s Largest Battery-Electric Ferry Sets Sail
The port of Buenos Aires, home terminal of ferry operator Buquebus, whose China Zorrilla now sails to Uruguay on battery power. Photo: Roberto Fiadone, via Wikimedia Commons, CC BY-SA 4.0.
Electric ferries are not new; small ones have crossed Norwegian fjords for years. What changed this month is scale. China Zorrilla, a 130-meter catamaran built by Incat Tasmania for operator Buquebus, has entered commercial service between Buenos Aires, Argentina, and Colonia, Uruguay. It is the largest ba...
More
Brown Engineers Beat the Dead Layer for Low-Power AI Chips
The crystal structure of bismuth ferrite (BiFeO3), the multiferroic material at the center of the Brown study. Image: Nan Wang and Ya Yang, via Wikimedia Commons, CC BY 4.0.
AI's appetite for electricity has become one of technology's defining problems. Better algorithms help, but some of the biggest long-term gains could come from rethinking the materials inside chips. Engineers at Brown University reported on October 28, 2026 that they have solved a long-standing obstacle for one promising...
More
A Laser Phase Plate Lets Cryo-EM See Proteins Too Small to Image
Illustration of a cryo-transmission electron microscope, the class of instrument the laser phase plate upgrades. Image: Ryan Kissinger, courtesy of NIAID, via Wikimedia Commons, public domain.
Cryo-electron microscopy (cryo-EM) has transformed how scientists see the molecules of life, earning its pioneers the 2017 Nobel Prize in Chemistry. Yet it has a stubborn blind spot: most human proteins are too small to image clearly. Physicists at the University of California, Berkeley, have now built ...
More
Oxford’s Bismuth Sulfide Finding Sidesteps a Key Solar Trade-off
Photovoltaic panels at the Travers Solar Farm in Alberta, Canada. Photo: Kkiefuik, via Wikimedia Commons, CC BY 4.0.
Every solar panel is fighting a quiet battle inside its materials. The better a semiconductor is at absorbing sunlight, the faster the electrical charges it creates tend to disappear before they can be collected. A study led by the University of Oxford and published in Nature on October 28, 2026 shows that this trade-off is not a law of nature. In a material called bismuth sul...
More
XREAL AURA: The First Android XR Glasses Put Gemini in View
XREAL's booth at Tokyo Game Show 2024, where the company showed earlier hardware; AURA is its first pair of glasses on Android XR. Photo: RuinDig/Yuki Uchida, via Wikimedia Commons, CC BY 4.0.
Headsets that cover your face have struggled to become everyday devices. Google and XREAL are betting that glasses will fare better. On October 28, 2026, the two companies launched XREAL AURA, the first wired XR glasses powered by Android XR. They project apps into your field of view while letting you ...
More
Apple’s MacBook Pro Goes OLED and Touch With PROLED XDR
A previous-generation 16-inch MacBook Pro (M2 Max), shown for reference; the new PROLED XDR models keep the familiar silhouette in a thinner, lighter body. Photo: SimonWaldherr, via Wikimedia Commons, CC BY-SA 4.0.
For more than a decade, Apple resisted putting a touchscreen on the Mac. On October 28, 2026, it changed course. The new MacBook Pro with PROLED XDR Display brings two firsts to Apple's pro laptop: a tandem OLED screen and touch input. It is one of the biggest redesigns of the Mac...
More
Humanoid Robots Scale Up: GXO Takes Agility’s Digit to More Sites
For years, humanoid robots have been more demo than deployment. That is why a quiet announcement on October 26, 2026 matters: GXO Logistics, the world's largest pure-play contract logistics provider, said it will expand its use of Agility Robotics' Digit humanoid to two more U.S. warehouses. A repeat order from a live customer is one of the clearest signals that a new technology is crossing from pilot to practice.
From one warehouse to three
According to the joint press release, Digit will ...
More
NVIDIA Nemotron 3.5 Transcribe: One Open Model for 40 Languages
A speech waveform, spectrogram and phonetic transcription, the raw material speech-recognition models learn from. Image: Wugapodes, via Wikimedia Commons, LGPL.
A transcript can capture every word and still miss what people need: who said what, a translation for a colleague, or the correct spelling of an industry term. On October 28, 2026, NVIDIA released Nemotron 3.5 Transcribe, a free, open-weight model designed to do all of that in one place, steered by plain-language instructions.
One...
More
