Atlas Energy Solutions Inc. (Atlas) has announced that it intends to expand its autonomous truck operations in the US Permian Basin from 28 to 100 by mid-2027. The company will work with Kodiak AI, Inc. (Kodiak) to make this ambition a reality using its physical artificial intelligence (AI)-powered autonomous driving technology.
What that means in practice is retrofitting or upfitting trucks with a package of sensors, controls, cameras, and other hardware to enable them to be operated autonomously using an AI “driver.”
According to Atlas, the new trucks will be used to expand operations at its second load-out point along its 42-mile (68 km) “Dune Express” sand conveyor system. This system will help Atlas become more responsive to customer needs for the transport and supply of proppant in the region.
Proppant, in case you are unaware, is a special grade of sand used by oil and gas companies during hydraulic fracturing. The sand is pumped underground to keep the cavities created by fracturing open, allowing oil and gas to flow freely.
Delivering sand without drivers
As you can appreciate, that requires enormous quantities of the stuff. What’s more, transporting it from mines or storage sites to drilling locations involves thousands of repetitive truck journeys, often on remote, dusty and relatively predictable routes.
The load-out points refer to strategic locations along the so-called “Dune Express” where this sand is loaded onto trucks. At present, Atlas explains, autonomous trucks have been used from one load-out point at a time.
The addition of another (one in Texas and one in New Mexico) will enable them to cover a wider area and improve customer order fulfillment. According to Atlas, since incorporating driverless trucks into its operations, they have carried approximately 7,000 loads totaling around 450,000 tons of sand over 23,500 “driverless hours.”
As for the trucks in question, they are entirely driverless but with an important distinction. To date, these trucks have tended to operate on controlled oilfield logistics networks, or private roads, not public highways or lanes.
That is considerably easier from an AI-training perspective than allowing autonomous trucks to travel freely through ordinary cities or mixed public traffic. The next major step is therefore operating on public roads, which Atlas and Kodiak say they expect to begin by early 2027, subject to regulatory and operational approval.
Fleet to increase to 100 by mid-2027
That would greatly expand where the trucks can deliver. “This partnership with Kodiak is a critical part of the Atlas strategy to transform oilfield sand logistics through innovation,” said John Turner, CEO and President of Atlas Energy Solutions, in a press release.
“Our mission is to seek critical energy infrastructure with inefficiencies, then engineer solutions that improve efficiency, reduce risk and enhance our customers’ operational success,” he added.
“Together, Atlas and Kodiak have deployed the world’s largest fleet of driverless big-rig trucks,” said Don Burnette, Founder and CEO, Kodiak.
“We’ve demonstrated that autonomous trucking is delivering meaningful results today. Our expanding partnership, which now covers two unique locations, is creating a blueprint for how autonomy can transform logistics by improving efficiency, increasing productivity, and delivering lasting value for customers at commercial scale,” he added.
Além de três modelos para programação, cibersegurança e agentes de IA, a Google revelou a mais recente versão da plataforma Gemini Robotics, que permite controlar o corpo inteiro dos robots e coordenar tarefas de forma mais autónoma.
Quadruped robots, which walk on four legs, are increasingly used for tasks such as inspection, transportation and search-and-rescue operations. However, their repeated leg movements consume far more energy than the rolling motion of wheeled robots, making it challenging to improve energy efficiency.
A teleoperated centaur robot dubbed “Threehalves” could represent the future of manual work in some of our planet’s most dangerous environments, according to Satyress Robotics, the California startup behind the project. In disasters such as mudslides, wildfires, and rubble searches, the company says its robot can keep workers safe from on-site hazards.
The other side of the story is the robot’s creepy, almost demonic appearance, as conveyed on the provocative Satyress Robotics website, the imagery for which is so unusual that it almost gives the impression of being a parody—particularly in an age of breathless reporting on corporate claims, and sudden appearances of “viral” news stories.
Therefore, The Debrief decided to take a deeper look at this mechanical oddity that more closely resembles something out of Christopher Nolan’s The Odyssey than a project in development in a California robotics lab.
An Unsettling Robot
Threehalves isn’t exactly news; all of the available information on the robot was posted on the company website over a year ago, around the time of their only post on X. Yet it’s recently been discovered by internet outlets noting the creepy design of this robot that the company claims is here to help.
“It’s obviously quite over the top, but ultimately, this is at the extreme end of humanoid robotics in size and capability,” a company representative told The Debrief.
“People may be familiar with humanoid robots from TV/movies or product videos, but likely don’t understand just how powerful and heavy they are (especially ours, which is coming in at just under 7ft tall and 240 lbs currently),” the rep added. “Our thought is someone who didn’t know what it is should definitely stay away from it, and those that do know or are operating the machine should be reminded it’s quite dangerous.”
“If you look at something like a lathe or mill, they generally look intimidating, for good reason, even though they could be dressed up to look friendly and soft,” the rep concluded. “We’ll point out, though, that the ‘evil’ look is mostly from it being pitch black; with different colors or a wrap on it, it does not put off that vibe.”
Robot “Superpowers”
Part of Threehalves’ sales pitch is that it won’t replace jobs, but instead will merely transfer them to robotic operators using a joystick to guide Threehalves through manual work in dangerous environments. The website claims that this amounts to granting workers “superpowers,” allowing them to focus on honing their craft in new extreme situations. However, while the operator won’t have to directly get their hands dirty, Satyress Robotics told The Debrief that Threehalves is designed for line-of-sight, close-proximity operation.
“We don’t have any plans to develop much autonomy above what’s needed to keep the robot from damaging itself,” the Satyress Robotics representative said. “Our goal is to have experienced trade experts in control of the bots at all times. Forestry, for instance, is the deadliest job in the US – our goal is to save lives first and foremost, and we believe that is also enough economically to justify the cost of the bot without any autonomous productivity gain.”
Two of the robots can fit in a standard truck bed, allowing them to be easily transported to remote locations. Lacking hands, the robot can be fitted with a variety of unique attachments to hold various tools from chainsaws to drills. Of note, the robot is shown holding a gas-powered chainsaw. Exactly how it would manage to get the chainsaw started without an obvious way to pull the ripcord is unclear.
Safety First
With fears of a demonic Terminator immediately conjured up by Threehalves’ appearance, the company informs us that although “this robot is more dangerous than it looks,” the same is true of any humanoid robot, and this one comes equipped with a variety of safety features.
Ironically, one of these is the robot’s menacing horns, which significantly slow the robot down in moving through standard-size door frames, and can even be replaced with larger horns that make traversing such portals impossible.
“It’s another way to make it clear this isn’t a toy, [a] security guard, etc,” the Satyress Robotics representative said. “It’s built for one reason, for a teleoperator to do dangerous work or heavy lifting from a safe, in line of sight distance.”
“We are not looking at any use cases in commercial buildings, and certainly not residential,” they added.
The recommended methods to halt the robot may be as disconcerting as its appearance. Credit: Satyress Robotics
A Violent Stop
Additionally, air reservoirs in the chest cavity and on the robot’s back are designed to activate pin locks on all its joints, rendering the robot immobile if air pressure is lost. Imagery on the website suggests this can seemingly be accomplished by stabbing Threehalves with a knife or shooting it with a gun. Fortunately, a more pacifist option to halt the robot also exists: simply by hitting an e-stop button located on its stomach.
When pressed on why the company would design their robot to be susceptible to gunfire specifically, the representative didn’t seem to find this the least bit unusual.
“Nothing software wise/control wise, whether autonomous or teleoperated, can be guaranteed to never have any kind of control issue,” the Satyress Robotics representative said. “This gives a means to lock the brakes remotely without necessarily damaging the robot beyond repair.”
Robot Reality or Hoax?
With the unsettling design and suggestion that the robot may eventually have to be shot to stop it from injuring a human, one of this reporter’s main questions became: “Is this a joke?” The outlandish nature of the project reminds one of prankster Joey Skaggs’ tricks on the media, such as the time he convinced WNBC-TV in New York and United Press International that he had recently developed a cure-all derived from roach hormones, which he was injecting into his followers.
However, the Satyress Robotics representative confirmed that the company’s new quadrupedal working robot is no joke.
“It’s not a prank. Also, we’re building it with the goal of keeping workers safe, so we just want to make it clear we’re thinking of every edge case and looking for ways to solve them outside of software,” the representative explained. “We are quite a ways away from showing the robot in any kind of elegant motion, thus the static renders/photos.”
Given that the company hasn’t posted on social media in over a year and has no LinkedIn presence, it was unclear who was behind the operation and if it was still underway. According to the company, they will shortly be making a video to introduce themselves.
While one picture on the website appears to show a live test with a joystick, there are no videos of Threehalves in motion on the website, and our email conversation confirmed it is not yet a functional robot.
For now, the company tells The Debrief that they are currently working on component-level testing and any commercial version of the Threehalves is more than two years away.
Ryan Whalen covers science and technology for The Debrief. He holds an MA in History and a Master of Library and Information Science with a certificate in Data Science. He can be contacted at ryan@thedebrief.org, and follow him on Twitter @mdntwvlf.
A Qualcomm exec suffered from one of the most disastrous tech demos we’ve ever seen live on stage at Computex 2026 this summer. In what should have been a wow moment during a press event in Taipei, Qualcomm hardware partner NEURA Robotics’ 4NE-1 humanoid robot quickly shifted gears from being a technical tour de force to a crashing mechanical shambles.
We see that the 4NE-1 humanoid robot is powered by Qualcomm’s Dragonwing IQ10 robotics reference design (RRD), claimed to be capable of 700 TOPS. But even before the robot’s spectacular malfunction, it didn’t look particularly healthy or nimble. 4NE-1’s slow shambling walking style radiated strong Two Soups vibes.
As the robot eventually hobbled close enough to the smart-suited presenter, the on-stage human picked up a Qualcomm Dragonwing IQ10 RRD from the robotically delivered platter and proudly showed it off to the attending crowd. As the applause and whoops subsided, the robot turned to face the crowd, and the presenter rolled on with his spiel.
Seconds into his script, the exec spun away to the side as the 4NE-1 humanoid robot very suddenly and noisily collapsed into a mechanoid mess. After resting in its terminal backbend for a second, one robot hand raised momentarily to make a futile gesture. Thankfully, the glass tray that had carried the Dragonwing IQ10 box onto the stage didn’t shatter.
We can think of another 'D' word for the strapline (Image credit: Qualcomm)
Stage hands seemed well-prepared
While the overall feeling from watching the video is one of shock from the ambling robot’s sudden noisy collapse, many on social media have commented on how unruffled and well-prepared the stagehands seemed to be. The silent and swift action of the black-clothed helpers, first cloaking the dead robot with purple silks, then the trio carrying it away, exits stage left, making it look like this catastrophe wasn’t such a surprise to some.
As the robotic remains were eventually removed through the side doors, we hear the brave speaker who carried on regardless come to the end of his polished presentation. The crowd politely and quite loudly claps - but is most of the applause for the speech, or for the unflappable stagehands who have efficiently extracted the heap that was the 4NE-1?
This late-emerging clip from Computex 2026 rivals the Windows 98 USB plug-and-play scanner crash during Bill Gates’ presentation at Comdex. That infamous software crash event may well have prompted the creation of the USB Cart of Death.
Legged robots have recently transitioned from science fiction to engineering fact, with modern humanoid and quadrupedal machines now capable of delivering packages to front doors and taking on dangerous military missions. With a massive surge in financial investment in the offing, a new study describes the technical advances that have made legged robots a reality and explores the critical ethical considerations, economic potential and policy implications of the "intelligent machines" that are increasingly walking among us.
University of Queensland researchers have developed new noninvasive sensors that measure muscle forces, unlocking new possibilities for wearable robotic mobility devices. Ultra-wideband radar sensors measure electromagnetic changes in muscles as they contract, allowing researchers to collect data in a way that's never been done before.
How do we combat the scourge of school shootings? The distinctly cyberpunk gamble that some places of learning are taking: rapid response drones.
Multiple schools across three states will soon have their own veritable platoons of pepper-spraying drones to hunt down active shooters, The Washington Post reports, waiting to spring into action the moment they’re called on.
Traveling at speeds of up to 60 miles per hour and capable of smashing through the windows, the so-called “Campus Guardian Angels” can reach a gunman within just fifteen seconds, the manufacturer Mithril Defense claims.
In addition to blasting a shooter with pepper spray, they can also distract them with loud sirens and flashing lights, and even ram into them at high speeds.
“Once I get the drones on you, you just can’t escape them,” Bill King, a Mithril cofounder and retired Navy SEAL, told WaPo.
It remains a fringe solution, but seems to be picking up steam. The state of Florida last year allocated more than $557,000 to a pilot program to deploy the drones at three different high schools. Georgia recently approved its own $500,000 pilot to test Mithril’s system at five public schools. And in Colorado, a new K-12 public charter school decided it would pay to use the drones on its own.
The drones aren’t autonomous and are instead remotely controlled by human operators at Mithril’s headquarters in Austin, Texas. Somewhere between 30 and 90 of the drones will be placed throughout a school in storage boxes, which helps them arrive quickly at the scene.
But the efficacy of the drone defense systems remains a big question mark — and hopefully we’ll never have to know — as is whether these solutions distract from addressing the systemic causes of mass shootings, like weak gun control and poor mental health support services.
Mithril has promised that its tech could help local law enforcement who respond to a shooting by locating the shooter and relaying critical information, but that would require a high degree of coordination in a hectic and dangerous environment. What if the drone operators target the wrong person, or cause law enforcement to mistake an innocent person for the shooter because the drones happen to be near them? Are these drone operators ready to make split-second decisions in a life or death situation for the people on the ground?
“It’s almost like where gaming and school safety collide,” Curtis Lavarello, CEO of the School Safety Advocacy Council, told WaPo. “I have some really serious reservations about the deployment of drones in schools.”
It’s not the only tech-inflected solution being adopted to combat mass shootings. There are already numerous startups offering AI-powered gun detection systems, but their performance has been dubious. One called Omnilert called the cops on an innocent 16-year-old after it mistook his bag of Dorito chips for a weapon, and is being sued by a survivor of a Tennessee high school shooting after it failed to detect the handgun used by the shooter.
Robotics company Realbotix LLC has long struggled to beat the allegations that its voluptuous humanoid robots are designed with prurient interests in mind.
Adding to the drama is the company’s corporate structure: Realbotix LLC is owned by Realbotix Corp, which also owns Intima LLC — a venture that holds an ownership stake in Realdoll, one of the best-known manufacturers of sex robots.
Those eyebrow-raising optics burst into view this month when, as New York Focus reports, the Salamanca City Central School District in western New York announced plans to deploy one of Realbotix LLC’s humanoid robots, dubbed Sally, in the classroom. The AI-enabled robot would reportedly serve as a teaching assistant for high school students enrolled in STEM and arts courses.
It’s hard not to see the company’s sex robot roots in Sally’s visage. The $57,000 android has smooth silicone skin, a dazed expression on her face, and an ill-fitting wig made of long brown hair. But after this story was first published, an attorney representing Realbotix reached out to emphasize what he said was a total divide between Realbotix LLC and Intima LLC.
“Realbotix Corp is a holding company with two separately operated businesses: Realbotix LLC, which focuses on AI-powered humanoid robotics for commercial service sector applications, and Intima LLC, which focuses on direct-to-consumer products,” the attorney said, in a statement echoing a press release the company issued last week. “Realbotix LLC and Intima LLC have different product offerings with no crossover. As a holding company, Realbotix Corp does not in itself have business operations but is responsible for managing capital markets, shareholders, listing requirements and accounting needs. The two subsidiaries maintain their own management, personnel, payroll, facilities, manufacturing operations, product development, and market strategies.”
“Realbotix LLC does not design, manufacture, or sell adult products,” the statement continued. “No Intima LLC or RealDoll employees, contractors, products, inventory, blueprints or technology are involved in Realbotix LLC’s design, engineering, manufacturing, deployment, or customer support whatsoever. All education, healthcare, wellness, and enterprise initiatives are conducted exclusively through Realbotix LLC. Realbotix LLC and Intima LLC are different legal entities that are independently operated. The two entities do not share employees, products, inventory, manufacturing resources, or technology.”
Prior to that reversal, district superintendent Mark Beehler told New York Focus that he was willing to give up banning AI in the classroom since “students will find a way around most rules that schools put in place.”
“I also believe it is critical for schools to teach proper use of technology, not simply exclude it,” he said.
Sally’s onboard AI can be used to generate lessons, provide translations in real-time in over 100 languages, and provide human teachers with guidance if a “teacher loses their place during a lesson or needs a prompt on what comes next,” according to New York Focus.
To us, that sounds like a pretty cynical take on the future of education. Teachers are already facing a major uphill battle thanks to the advent of the tech, with students suffering from declining cognitive abilities.
Of course, Realbotix LLC argues that it’s done its homework. Instead of hallucinating facts, a major shortcoming of the tech that makes it arguably unfit for most classroom use, Sally will allegedly say “I don’t know.” The company also said its robot won’t be directly connected to the internet to keep highly sensitive information about students from prying eyes.
Still, it’s a bizarre investment for the school district that comes at a particularly tough time for human educators. The reality is that teachers across the country are woefully underpaid. There’s a massive shortage of instructors, with tens of thousands of teacher positions going unfilled.
In a press release, the Salamanca City Central School District stressed that the “Realbotix educational robot will never replace teachers, staff members, or meaningful human interaction.”
But whether Sally will actually serve as “another instructional tool that educators can use to engage students, reinforce classroom learning, and spark curiosity about emerging technologies,” remains to be seen — especially considering its steep price tag and lascivious origins.
Updatedto clarifythe relationship between Realbotix Corp, Realbotix LLC, and Intima LLC.
With so many landmark lawsuits emerging around artificial intelligence, it can be easy to lose sight of the other high-tech cases shaping the world as we know it.
One major suit with important implications just emerged out of Phoenix. The case revolves around a 2023 incident involving a 73-year-old named Trudy Perez and a delivery robot at Arizona State University. The woman — who was working as a parking garage attendant at the time — alleges a delivery robot operated by the company Starship Technologies repeatedly ran into her, throwing her to the ground and clobbering her again while she was down.
Now, the woman is proceeding with a civil suit targeting Starship for liability and negligence, local station KJZZ reported.
As if working outdoors in Phoenix at the age of 73 isn’t rough enough, a police report from the incident notes Perez suffered a four-inch gash on her arm as a result of the beat-down, in addition to back pain and mobility problems. Per KJZZ, she’s now seeking compensatory damages related to medical expenses, pain and suffering, and lost wages caused by the altercation.
The lawsuit further alleges that, like similarautomatedrobots, the experimental Starship unit was being piloted by an off-site employee. Per the filing, this unnamed employee neglected to “exercise control” over the delivery bot throughout the incident, resulting in Perez’s injuries.
Of course, the company is doing everything it can to soften the blow and limit its liability. As KJZZ reports, Starship Technologies has a laundry list of excuses: not only were Perez’s injuries caused by her own negligence, but they were pre-existing as well, the company’s attorneys claim.
With the first court date set for next February, it’ll be some time before the jury reaches a verdict. When they do, they could set some major precedent when it comes to holding companies responsible formayhem caused by their delivery robots — an increasingly important issue as the gadgets spread to more cities across the US.
Despite fears that China could race ahead of the US in AI, a joint UK-US evaluation found that Moonshot AI’s Kimi K3 performs “significantly below” leading American models in offensive cybersecurity capabilities.
The findings address growing concerns in Washington regarding China’s recent advancements in artificial intelligence following the release of Kimi K3, described as China’s most powerful large language model.
The UK Artificial Intelligence Security Institute (UK AISI) and the US Center for AI Standards and Innovation (CAISI) found that Kimi K3 achieved an overall cyber capability score of 32.2%, compared to an average of 76.2% for top, unnamed US models.
Evaluation identifies cybersecurity gap
Researchers tested Kimi K3 using ExploitBench, a public benchmark developed by Carnegie Mellon University that measures an AI system’s ability to develop end-to-end exploits for software vulnerabilities.
According to the evaluation, Kimi K3 “performs significantly below the most recent frontier cyber-capable models” when tasked with developing exploits. The model attained a score of 32 percent, outperforming GLM-5.2, another Chinese open-weight model, which scored 24 percent.
However, Kimi K3 still lagged behind leading US systems. The most cyber-capable US models achieved an average score of 76.2%.
The report also examined whether models could achieve arbitrary code execution (ACE)—a high-severity exploit outcome that allows attackers to take control of a target system. Kimi K3 achieved ACE on none of the 41 ExploitBench tasks, while “the most cyber-capable models achieved ACE on 20/41 samples on average,” according to the UK AISI and CAISI.
Simulated attack shows remaining capabilities
The evaluation also tested Kimi K3 on “The Last Ones” (TLO), a 32-step simulated corporate network attack designed to measure a model’s ability to conduct end-to-end cyber operations autonomously.
Kimi K3 reached step 17 of the attack path on average, compared with the 28.5 steps achieved by the most cyber-capable US models.
Despite this gap, the report noted that Kimi K3 completed the simulated attack in one of 10 attempts. The UK AISI and CAISI said this demonstrated that the model “is capable of autonomously attacking small, weakly defended, and vulnerable enterprise systems, when directed to do so and given initial network access.”
Furthermore, Kimi K3’s safeguards did not prevent it from attempting cyber exploit development or offensive cyber operations during testing.
Although the results were only preliminary, the research noted that Kimi K3’s overall cyber capability score was estimated from a single benchmark, whereas other models were evaluated across a larger number of cyber tasks.
Findings reshape AI competition debate
Kimi K3 has sparked attention, if not concern, following reports of its strong performance on other AI benchmarks. However, these new findings offer a different picture of the model’s cybersecurity capabilities, showing that it still lags behind US systems.
The South China Morning Post reported that the evaluation “offers a stark reality check amid mounting panic in Washington over Beijing’s AI advancements following Moonshot AI’s latest flagship release.”
David Sacks, former White House AI czar and current co-chair of US President Donald Trump’s Council of Advisers on Science and Technology, said the findings suggest that fears over Chinese model supremacy may be exaggerated. He urged policymakers to avoid over-regulating the sector.
“The Kimi Panic needs to stop,” Sacks said in a social media post. “American frontier models are still ahead.”
US firm Rescale, a digital engineering platform that combines AI, simulation, and high-performance computing to help R&D and engineering teams design, test, and bring new products to market faster, has received funding under the Genesis Mission, a national initiative led by the US Department of Energy to build an advanced integrated science discovery platform.
As part of the program, Rescale will join Lawrence Berkeley National Laboratory, Lawrence Livermore National Laboratory, and Oak Ridge National Laboratory as a funded partner in the Agentic HPC Pipeline Initiative (AHPI).
The initiative aims to use agentic AI on Rescale’s platform to help accelerate the adoption of national laboratory-developed computing tools and simulations by US industry.
Doubling scientific research output with AI and advanced computing
The Genesis Mission is an ambitious initiative aimed at accelerating scientific and technological breakthroughs by combining artificial intelligence, supercomputing, quantum systems, and advanced scientific instruments into a single research platform.
Established through an executive order in November 2025, the program brings together the department’s 17 national laboratories and around 40,000 scientists and engineers to advance research in areas including energy, scientific discovery, and national security.
Department officials have compared the scale and ambition of the initiative to landmark US efforts such as the Manhattan Project and the Apollo program, with the goal of doubling the productivity of the country’s scientific research within the next decade.
The US Department of Energy launched the first Genesis Mission funding round earlier this year, awarding $293 million to projects focused on areas including advanced manufacturing, biotechnology, nuclear energy, and quantum science. The program received more than 8,000 applications, making it the department’s largest-ever funding call.
Rescale’s AHPI was selected for Phase I funding to address a key challenge: helping industry access advanced simulation tools developed by US national laboratories that have often required specialized expertise and infrastructure.
The company will work with three US national laboratories to bring agentic AI capabilities to advanced simulation tools. The project will integrate AI agents into WarpX, LiDO, and Adamantine, codes developed by Lawrence Berkeley National Laboratory, Lawrence Livermore National Laboratory, and Oak Ridge National Laboratory.
Using the Rescale platform, AI agents will help engineers choose simulation tools, configure inputs, monitor runs, and analyze results. The initiative aims to reduce the time and specialized expertise needed to use these simulations by more than five times, making advanced research capabilities more accessible to US manufacturers.
According to Rescale CEO Joris Poort, the Genesis Mission highlights the importance of making industrial AI technologies available to leading research and development teams that can apply them in real-world applications.
Building on Rescale’s existing partnership with ORNL’s Manufacturing Demonstration Facility, AHPI addresses a common gap: DOE-industry collaborations run for a fixed term, and once they end, partners keep the insight but lose the tools to act on it. AHPI keeps the resulting agentic workflows, surrogate models, and code environments live on Rescale, letting labs and industry keep collaborating long after a project’s funding ends.
Rescue teams in coastal cities and towns often need to rescue people who are at risk of drowning. These operations often require emergency teams to locate and reach people in distress as quickly as possible, as even a short delay could have serious or even fatal consequences.
Robots slithered like snakes into the rubble of collapsed buildings to search for trapped survivors in the aftermath of Venezuela’s double earthquake disaster.
The so-called snake robots represented a unique technological aid for international search and rescue efforts following two earthquakes that struck Venezuela on June 24, killing more than 5,000 people and leaving behind 16,740 injured. Rescuers from around the world used trained dogs, seismic and acoustic listening devices, video cameras on sticks, or thermal imaging cameras to search for signs of survivors—but the snakebots equipped with a video camera could squeeze through tight spaces to search deep within fallen building debris.
“The reason why snake robots are useful is they can get into tightly packed volumes that machinery and people otherwise cannot, and that's important in search and rescue because we were able to extend the visual reach of the rescue workers in a safe way,” Howie Choset, head of the Biorobotics Lab at Carnegie Mellon University in Pittsburgh, told Ars. “In a sense, you're doing minimally invasive surgery on a structure with a snake robot.”
Robots slithered like snakes into the rubble of collapsed buildings to search for trapped survivors in the aftermath of Venezuela’s double earthquake disaster.
The so-called snake robots represented a unique technological aid for international search and rescue efforts following two earthquakes that struck Venezuela on June 24, killing more than 5,000 people and leaving behind 16,740 injured. Rescuers from around the world used trained dogs, seismic and acoustic listening devices, video cameras on sticks, or thermal imaging cameras to search for signs of survivors—but the snakebots equipped with a video camera could squeeze through tight spaces to search deep within fallen building debris.
“The reason why snake robots are useful is they can get into tightly packed volumes that machinery and people otherwise cannot, and that's important in search and rescue because we were able to extend the visual reach of the rescue workers in a safe way,” Howie Choset, head of the Biorobotics Lab at Carnegie Mellon University in Pittsburgh, told Ars. “In a sense, you're doing minimally invasive surgery on a structure with a snake robot.”
Professor Sukho Song of the Department of Robotics and Mechatronics Engineering at DGIST has proposed "Sustainability Robotics," a new academic field that evaluates robots not only in terms of their technical performance but also their contributions to environmental, social and economic sustainability.
Engineers have created a drone that camouflages itself as a faint, hovering haze when it takes flight.
The "Phantom Twist" prototype drone is a carefully engineered robot that utilizes motion blur to hide in plain sight — opening up a new frontier in low-visibility flight. Scientists say it could reshape how drones monitor wildlife, survey ecosystems and inspect infrastructure.
Most stealth drones try to disappear by changing their appearance, using camouflage paint, transparent plastics or complex optical tricks to bend or scatter light. The Phantom Twist takes a fundamentally different approach by exploiting how humans and animals perceive motion.
The scientists who created the drone outlined their findings in a study presented July 16 at the conference Robotics: Science and Systems in Sydney, Australia.
"Most efforts to hide drones focus on making them look like their surroundings," said first author of the study Michael Rubenstein, associate professor of computing and mechanical engineering at Northwestern University, in a statement. "Instead, we asked whether we could design the drone itself around the way humans perceive motion.”
(Image credit: Michael Rubenstein/Northwestern University)
The result is a machine that behaves more like a spinning fan than a conventional quadcopter. Typical drones have four rotors spinning around a largely stationary body, so your eye can lock onto the frame even as the propellers blur.
Phantom Twist strips this idea down to a single motor and single propeller. The propeller spins one way while the entire body counter‑rotates in the opposite direction with no stationary parts to visually latch on to — spinning 25 times per second, which is too fast for the human eye to resolve clearly.
(Image credit: Michael Rubenstein/Northwestern University)
It's about 10 times less visually perceptible than a conventional quadcopter, the scientists said. In practice, that means observers don’t see a crisp flying machine but a translucent smudge, with its few opaque components blending into a haze.
To get there, the team assigned algorithms to explore thousands of designs. First, they used a computational model to automatically generate about 20,000 drone configurations that could fly stably, at least in theory. Each configuration shuffled a motor, propeller, batteries, circuit board and counterweight into different positions.
Then artificial intelligence (AI) and optimization tools took over, iteratively rearranging those pieces to minimize how visible each design would be from almost any viewing angle — all while respecting the constraints of aerodynamic stability.
The researchers simulated each candidate spinning in midair and composited the images over 100 different real-world backgrounds, from skies to trees to buildings.
To determine the stealthiest designs, they applied a perception model that approximated human vision, with drones that better blended into their surroundings earning a lower visibility score. From there, the team selected the 500 lowest-scoring designs and ran their optimization loop again, squeezing out further decreases in visibility. When the team was confident in the design of a particular drone, they built it.
(Image credit: Michael Rubenstein/Northwestern University)
The final Phantom Twist isn’t a compact block of hardware; it’s deliberately spread out in three dimensions. Components sit at different heights and angles, with empty space between them, so that when the drone spins, the parts don’t visually overlap into a solid silhouette.
That geometry matters because of the way our eyes process images.
"The human eye takes time to accumulate signals, roughly analogous to the exposure time of a camera," said computer vision expert and co-author of the study Emma Alexander. When something spins fast enough, distinct edges smear together, and we perceive a blur instead of a shape.
The researchers' initial uses for the stealthy drones include monitoring nesting birds without startling them, surveying wetlands without scattering flocks of waterbirds, and inspecting aging infrastructure without altering human behavior, they said.
For now, Phantom Twist isn’t a perfect "ghost." The spinning propeller still generates a significant amount of noise and the spider‑web of thin wires and support rods remains partially visible. Rubenstein and his colleagues said in the study they’re already thinking about next-generation iterations that swap in more transparent materials and quieter propulsion systems.
We can think of no possible way that a newly invented invisible drone can be used nefariously.
Rather than relying on camouflage, the flying gizmo created by researchers at Northwestern University exploits the motion blur effect in our fallible human vision to almost disappear by spinning at an absurd 25 times per second.
This is too quick for us and many animals to see clearly, becoming a gossamer smear that blends into the background, like the way the blades of a helicopter do, only more extreme.
One day, this type of design could be used for completely peaceful pursuits, the researchers say — like observing wildlife, of course.
“Most efforts to hide drones focus on making them look like their surroundings,” Michael Rubenstein, an associate professor of computer science and mechanical engineering at Northwestern who led the research, said in a statement about the work.
“Instead, we asked whether we could design the drone itself around the way humans perceive motion,” he added. “This idea of low visibility through persistent motion is something few people have explored.”
To create the drone, dubbed “Phantom Twist,” the team used a computational model to churn out around 20,000 different drone designs, then used AI and optimization algorithms to repeatedly rearrange how its major parts, like their motor and propeller, were configured.
Then they took the best configurations and used a model that simulates human perception to quantify how visible each design appeared. The least visible designs were chosen to be further rearranged with the optimization algorithm, until landing on the blurriest candidate of them all.
“The design process was fully automated,” Rubenstein said. “Then, when we were confident that a drone met all our criteria, we built it.”
It’s not the first spinning drone, but it is the first that’s been optimized to be effectively invisible to the human eye. Its major components are carefully spaced enough from each other to never become a distinct shape when viewed from any angle while spinning — all while being capable of stable flight.
According to the researcher’s visibility model, this allows the Phantom Twist to be about ten times less visually perceptible than a conventional quadcopter drone. Impressive footage certainly bears this out: after the drone gets up to full speed, it almost looks like the ghost of a tiny UFO.
The drone is quite noisy, however. If something like it is going to be used to observe wildlife one day, it’ll have to be vastly quieter. Until then, if these faint apparitions are ever turned on humans, at least we’ll be able to hear them coming.
As AI models continue to grow in size and complexity, chipmakers are shifting their focus from individual processor performance to the ability to efficiently connect thousands of accelerators into a single computing system.
These highly interconnected server architectures have become a key competitive area for AI infrastructure companies, which are racing to deliver the computing scale needed for models with trillions of parameters and the growing adoption of AI agents.
According to Ding Yunfan, vice president of AI framework architecture at Biren, improvements in the performance of a single graphics processing unit (GPU) are no longer enough to meet these demands. Instead, the industry’s biggest challenge is turning vast GPU clusters into unified, seamless computing platforms capable of operating as a single system.
Architecture designed to break GPU cluster limits
Biren believes the future of large-scale AI computing depends on replacing conventional electrical interconnects with optical technology. According to Ding, traditional copper-based connections are approaching their physical limits, effectively capping today’s server architectures at around 128 GPUs, the South China Morning Postreports.
To overcome that bottleneck, the company has developed a distributed, decoupled supernode architecture built around near-packaged optics (NPO), which positions optical fibres closer to the chips to increase bandwidth and improve data transmission. Biren said the approach could enable clusters with up to 1,024 AI accelerator cards, significantly expanding the scale of computing systems needed to support increasingly demanding AI workloads.
Ding said optical interconnects are becoming essential for overcoming the scalability limits of conventional server architectures. Shanghai-based Biren is among a growing number of Chinese companies tackling these infrastructure challenges. Rivals including MetaX, Enflame and Alibaba are also developing supernode architectures to compete with Nvidia’s NVL72 and next-generation NVL144 AI platforms.
Biren’s Shanghai-based rival MetaX has introduced its Xijing S600 AI supernode, featuring a single rack that connects 64 GPUs. The company said its design reduces signal loss by removing external cabling between compute nodes, switch nodes and interconnect systems, improving efficiency for large-scale AI workloads.
Looking beyond electrical links to build larger AI supernodes
Biren is also working on new approaches to expand the scale of its AI computing systems. The company’s orthogonal hardware architecture, co-developed with telecommunications giant ZTE, uses traditional electrical connections, while Biren is simultaneously testing a prototype near-packaged optics (NPO) interconnect system.
The chipmaker said the optical approach could eventually support supernodes with more than 512 accelerator cards, helping address the growing demand for larger and more efficient AI infrastructure.
Despite growing interest in optical networking for AI infrastructure, Ding said the semiconductor industry is still in the early stages of developing and testing optical interconnect technologies. He noted that widespread adoption will require further advances before these systems can be deployed at scale.
Furthermore, Ding emphasized that the technology still requires real-world validation before it can be widely adopted. He projected that NPO optical systems could reach mass commercial deployment around 2028, as the industry continues refining the technology and addressing remaining engineering challenges.
Login
