It’s easy to assume that Robert Woo was defined by the accident that took away his ability to walk.
Certainly, the day of his accident—14 December 2007—was a turning point. Woo, an architect working on the new Goldman Sachs headquarters in New York City, hadn’t attended his company’s holiday party the night before, and that morning he was the only one in the trailer that served as the construction-site office. He was bent over his laptop when, 30 floors above, a crane’s nylon sling gave way, sending about 6 tonnes of steel plummeting toward the trailer. The roof collapsed, folding Woo in half and smashing his face into his laptop, which smashed through his desk.
“I was conscious throughout the whole ordeal,” Woo remembers. “It was an out-of-body experience. I could hear myself screaming in pain. I could hear the voices of the rescue workers. I heard one firefighter say, ‘Don’t worry, we’re getting to you.’” The rescue workers hauled him out of the rubble and got him to the emergency room in 18 minutes flat; with one lung crushed and the other punctured, he wouldn’t have lasted much longer. In those frantic early moments, a doctor told him that he might be paralyzed from the neck down for the rest of his life. He remembers asking the doctors to let him die.
Woo simply couldn’t imagine how a paralyzed version of himself could continue living his life. Then 39 years old, he worked long hours and jetted around the world to supervise the construction of skyscrapers. More important, he had two young boys, ages 6 months and 2 years. “I couldn’t see having a life while being paralyzed from the neck down, not being able to teach my boys how to play ball,” he recalls. “What kind of life would that be?”
Advertisement
Robert Woo walks inside the Wandercraft facility in New York City using the company’s latest self-balancing exoskeleton. Nicole Millman
But in a Manhattan showroom last May, Woo showed that he’s not defined by that accident, which left him paralyzed from the chest down, but with the use of his arms. Instead, he has defined himself by how he has responded to his injury, and the new life he built after it.
In the showroom, Woo transferred himself from his wheelchair to a 80-kilogram (176-pound) exoskeleton suit. After strapping himself in, he manipulated a joystick in his left hand to rise from a chair and then proceeded to walk across the room on robotic legs. Woo’s steps were short but smooth, and he clanked as he walked.
This exoskeleton, from the French company Wandercraft, is one of the first to let the user walk without arm braces or crutches, which most other models require to stabilize the user’s upper body. The battery-powered exoskeleton took care of both propulsion and balance; Woo just had to steer. The bulky apparatus had a backplate that extended above Woo’s head, a large padded collar, armrests, motorized legs, and footplates. Walking across the room, he appeared to be half man, half machine. On the other side of the showroom’s plate-glass window, on Park Avenue, a kid walking by with his family came to a dead halt on the sidewalk, staring with awe at the cyborg inside.
Advertisement
Robert Woo prepares to walk in a Wandercraft exoskeleton; the device’s controller enables him to stand up, initiate walk mode, and choose a direction. Bryan Anselm/Redux
The amazement on the boy’s face was reminiscent of Woo’s young sons’ reaction when they saw a photo of Woo trying out an early exoskeleton, back in 2011. “Their first comment was, ‘Oh, Daddy’s in an Iron Man suit,’” he remembers. Then they asked, “When are you going to start flying?” To which Woo replied, “Well, I’ve got to learn how to walk first.”
The title of exoskeleton superhero suits Woo. He’s as soft-spoken and mild-mannered as Clark Kent, with a smile that lights up his face. Yet the strength underneath is undeniable; he has built a new life out of sheer determination.
For 15 years, he’s been a test pilot, early adopter, and clinical-study subject for the most prominent exoskeletons under development around the world. He placed the first order for an exoskeleton that was approved for home use, and he learned what it was like to be Iron Man around the house. Throughout it all, he has given the companies detailed feedback drawn from both his architectural design skills and his user experience. He has shaped the technology from inside of it.
Saikat Pal, a researcher at the New Jersey Institute of Technology, in Newark, met Woo during clinical trials for Wandercraft’s first model. Like so many others in the field, Pal quickly recognized that Woo brought a lot to the table. “He’s a super-mega user of exoskeletons: very enthusiastic, very athletic,” Pal says. “He’s the perfect subject.”
Advertisement
By pushing the technology forward, Woo has paved the way for thousands of people with spinal cord injuries as well as other forms of paralysis, who are now benefiting from exoskeletons in rehab clinics and in their homes. “Our bionics program at Mount Sinai started with Robert Woo,” says Angela Riccobono, the director of rehabilitation neuropsychology at Mount Sinai Hospital, in New York City, where Woo became an outpatient after his accident. “We have a plaque that dedicates our bionics program to him.”
Robert Woo walks down a sidewalk in New York City in 2015 using a ReWalk exoskeleton, one of the first exoskeletons designed for use outside the rehab clinic. Eliza Strickland
It’s a fitting tribute. Woo’s post-accident life has been marked by victories, frustrations, deep love, and one devastating loss, and yet he has continued to devote himself to bionics. And while his vision for exoskeletons hasn’t changed, experience has reshaped what he expects from them in his lifetime.
Long before Woo ever stood up in a robotic suit, he had developed the habits of mind that would later make him an unusually perceptive test pilot.
Advertisement
Woo has always been a builder, a tinkerer, a fixer. Growing up in the suburbs of Toronto, he put together model kits of battleships and airplanes without looking at the instructions. “I just put things together the way I thought it would work out,” he says. He trained as an architect and in 2000 joined the Toronto-based firm Adamson Associates Architects, a job that soon had him traveling to Europe and Asia to work on corporate high-rises.
Adamson specializes in taking the stunning designs of visionary architects and turning them into practical buildings with elevators and bathrooms. “Most of the design architects don’t really have a clue about how to build buildings,” Woo says. He liked solving those problems; he liked reconciling beautiful designs with the stubborn reality of construction. That talent for understanding a structure from the inside and spotting the flaws would prove essential later.
After his accident, Woo had two major surgeries to stabilize his crushed spine, which required surgeons to cut through muscles and nerves that connected to his arms. For two months, he couldn’t feel or move his arms; there was a chance he never would again. Only when sensation began creeping back into his fingertips did he allow himself to imagine a different future. If he wasn’t paralyzed from the neck down, he thought, maybe more of his body could be brought back online. “My focus was to walk again,” he says.
Woo was discharged in March 2008 and went back to his New York City apartment. He was still bedridden and required around-the-clock care. He doesn’t much like to talk about this next part: By May, his then-wife had moved back to Canada and filed for divorce, asking for full custody of their two children. Woo remembers her saying, “I can’t look after three babies, and one of them for life.”
Advertisement
It was a dark time. Riccobono of Mount Sinai, who met Woo shortly after he became an outpatient there in 2008, recalls the despondent look on his face the first time they talked. “I wasn’t sure that he wasn’t going to take his life, to be honest,” she says. “He felt like he had nothing to live for.”
Angela Riccobono of Mount Sinai Hospital (left) credits Woo with jump-starting the hospital’s bionics program; a plaque in the department of rehabilitation medicine recognizes his role.
Yet Woo harbors no animosity toward his ex-wife. “If we hadn’t separated and gone through the custody hearing, I don’t think I would have gotten this far,” he says. To win partial custody of his children, Woo had to become independent. He had to get off narcotic pain medications, regain strength, and learn how to navigate life in a wheelchair. He had to show that he no longer needed constant nursing, and that he could take care of both himself and his boys.
There were milestones: learning how to get back into his wheelchair after a fall, learning to drive a car with hand controls, learning to manage his body as it was, not as it had been. The biggest change came when he reconnected with his high school sweetheart, a vivacious woman named Vivian Springer. She was then dividing her time between Toronto and New York City, and she had a son who was almost the same age as Woo’s two boys. Springer had worked in a nursing home and knew how to change the sheets without getting him out of bed; she was currently working in human resources and knew how to deal with insurance companies. “You wouldn’t believe how much stress it lifted off of me,” Woo says. Over time, they became a family.
Robert Woo’s wife, Vivian, was trained in how to operate the device he used at home. His sons, Tristan (left) and Adrien, grew up watching their dad test exoskeletons. Left: Lifeward; Right: Robert Woo
Once Woo had that foundation in place, Riccobono witnessed a profound change. “He went from focusing on ‘what I can’t do anymore’ to ‘What’s still possible? What can I do with what I have?’” At Mount Sinai, Woo remembers asking his doctor Kristjan Ragnarsson, who was then chairman of the department of rehabilitation medicine, if he would ever walk again. “His response was, ‘Yes, you can walk again,’” Woo remembers, “‘but not the way you used to walk.’”
Advertisement
First Steps in an Exoskeleton
As soon as he had regained use of his hands, Woo had started googling, looking for anything that could get him back on his feet. He tried rehab equipment like the Lokomat, which used a harness suspended above a treadmill to enable users to walk. But at the time, it required three physical therapists: one to move each leg and one to control the machine. It was a far cry from the independent strides he dreamed of.
Several years in, he learned about two companies that had built something radically different: exoskeleton suits for people with spinal cord injuries. These prototypes had motors at the knees and the hips to move the legs, with the user stabilizing their upper body with arm braces. Woo desperately wanted to try one, although the technology was still experimental and far from regulatory approval. So he took the idea to Ragnarsson, asking if Mount Sinai could bring an exoskeleton into its rehab clinic for a test drive. Ragnarsson, who’s now retired, remembers the request well. “He certainly gave us the kick in the behind to get going with the technology,” he says.
Robert Woo tries out an early exoskeleton from Ekso Bionics at Mount Sinai Hospital, where he first began testing the technology. Mario Tama/Getty Images
Ragnarsson had seen decades of failed attempts to get paraplegics upright, including “inflatable garments made of the same material the astronauts used when they went to the moon,” he says. All those devices had proved too tiring for the user; in contrast, the battery-powered exoskeletons promised to do most of the work. And he knew one of the founders of Ekso Bionics, a Berkeley, Calif.–based company that had built exoskeletons for the military. In 2011, Ekso brought its new clinical prototype to Mount Sinai.
The day came for Woo’s first walk. “I was excited, and I was also scared, because I hadn’t stood up for almost five years,” he remembers. “Standing up for the first time was like floating, because I couldn’t feel my feet.” In that first Ekso model, Woo didn’t control when he stepped forward; instead, he shifted his weight in preparation, and then a physical therapist used a remote control to trigger the step. Woo walked slowly across the room, using a walker to stabilize his upper body, his steps a symphony of clunks and creaks and whirs. He found it mentally and physically exhausting, but the effort felt like progress.
Advertisement
Robert Woo stands using an exoskeleton and embraces his wife, Vivian. Woo says that exoskeleton use has both physical and psychological benefits. Mt. Sinai
Riccobono was there for those first steps, with tears running down her face. “I remembered how he looked the day I first met him, so defeated,” she says. “To see him rise from the chair, to see him rise to a standing position, to see how tall he was, to see him take those first steps—it was beautiful.” Ragnarsson saw clear benefits to the technology. “Any type of walking is good physiologically,” he says. “And it’s a tremendous boost psychologically to stand up and look someone in the eye.” Woo remembers hugging his partner, Springer, and for the first time not worrying about running over her toes with his wheelchair. I first met Woo a few days later, during his third session with the Ekso at Mount Sinai.
Ann Spungen (left), a researcher at a Veterans Affairs hospital, led early clinical trials of exoskeletons. Her research focused on the medical benefits of exoskeleton use. Robert Woo
Later that same year, at a Department of Veterans Affairs (VA) hospital in the Bronx, Woo got to try a prototype of the world’s other leading exoskeleton: the ReWalk, from the Israeli company of the same name (since renamed Lifeward). VA researchers, led by Ann Spungen, were keen to determine if exoskeleton use had real medical value for veterans with spinal cord injuries. Woo was part of that clinical trial, for which he had more than 70 walking sessions, and he’s since been in many others. But he remembers the first VA trial with the most gratitude. “Dr. Spungen’s first exoskeleton clinical trial really turned things around for me,” he says.
Over the course of the trial’s nine intense months, Woo says he saw noticeable improvements to many facets of his health. “By the end of the trial, I eliminated about three-quarters of my medication intake,” he says, including narcotic pain pills and medication for muscle spasms. He grew fitter, with less body fat, more muscle mass, and lower cholesterol. His circulation improved, he says, causing scrapes and cuts to heal more quickly, and his digestion improved too. The results Woo experienced have generally been borne out in research studies at the VA and elsewhere—exoskeletons aren’t just good for the mind, they’re good for the body.
Advertisement
Improving Exoskeletons From the Inside
During the VA trial, Woo began to think of exoskeletons not as miraculous machines, but as works in progress.
Pierre Asselin (right), a biomedical engineer, worked with Robert Woo during clinical trials of exoskeletons. He says Woo was always pushing the limits of the technology. Robert Woo
Pierre Asselin, the biomedical engineer coordinating the VA’s study, watched participants respond very differently to the equipment. “These devices are not the equivalent of walking—you’re tired after walking a mile,” he says. He notes that later models of both the Ekso and ReWalk enabled users to initiate each step through software that recognized when they shifted their weight. Asselin adds that the cognitive load is “like learning to drive a manual transmission car, where at first you’re really struggling to coordinate the clutch and the brake.” Woo picked it up immediately, he remembers.
Robert Woo uses an exoskeleton to reach items in a kitchen cabinet during a test of the device’s utility for everyday tasks. Eliza Strickland
Woo became an invaluable partner, Asselin says. “When we first started with the devices, there was no training manual. We developed all of that through collaboration with Robert and other participants.” Woo pushed the limits of the technology, Asselin says, whether it was seeing how many steps he could take on one battery charge or simulating a failure mode. “He’d say, ‘What happens if I was to fall? What would be the approach to getting up?’”
Woo approached the ReWalk the way he had approached buildings in his previous life: He looked inside the structure and found the weak points. An early model left some users with leg abrasions where the straps rubbed—a small injury for most people, but a serious risk for someone who can’t feel a wound forming. Woo suggested better padding and stronger abdominal supports to redistribute the load. He also hated the heavy backpack that carried the battery and computer, so one afternoon he grabbed an old pack, cut off the straps, and rebuilt it into a compact hip-mounted pouch. Then he snapped photos and sent them to the company. The next model arrived with a fanny pack.
Advertisement
Robert Woo sent detailed design sketches as part of his feedback to exoskeleton engineers. Robert Woo
Sometimes his fixes were more ambitious. One Ekso unit that he used at Mount Sinai kept shutting down after 30 minutes. Woo felt the hip motors and found them hot to the touch. “I said, ‘Can I remove these? I’m going to make a really quick fix, okay? Give me a drill and I’ll put a couple of holes in it,” he recalls telling the therapists, proposing to create a DIY heat sink. He wasn’t allowed to modify the prototype, but a year later the company introduced improved cooling around the hip motors. “There is a Robert Woo design on this device,” one therapist told him.
Eythor Bender, who was then the CEO of Ekso, called Woo to thank him for his feedback and invite him to spend a week at Ekso’s headquarters. “There was no lack of engineering power in that building,” says Bender. “But sometimes when you work with engineers, they overlook important things.” Bender says Woo brought both design skills and lived experience to his weeklong residency. “He told the engineers, ‘Guys, this has to be something that people actually like to wear.’”
Ekso Bionics CEO Eythor Bender and Mount Sinai physician Kristjan Ragnarsson were both on hand for Woo’s early trials of the Ekso device. Ragnarsson says he saw physical and psychological benefits of exoskeleton use. Robert Woo
The longer Woo tested, the further ahead he started thinking. With motors only at the hips and knees, every exoskeleton still required crutches. Add powered ankles, he told the Ekso and ReWalk teams, and the suits could balance themselves, freeing the user’s hands. But Woo was ahead of his time. “They said they weren’t going to do that. They weren’t going to change their whole platform,” he remembers. Years later, though, hands-free exoskeletons like those from Wandercraft would emerge built around exactly that principle.
When the Exoskeleton Came Home
By the mid-2010s, Woo had pushed the technology as far as he could in clinics. What he wanted now was to use an exoskeleton at home.
Advertisement
That milestone came after ReWalk’s exoskeleton became the first to win FDA approval for home use in 2014. ReWalk engineers still remember Woo’s help on the final tests for that personal-use model. It was the end of May in 2015, recalls David Hexner, the company’s vice president of research and development. “He said, ‘Guys, this is great. I’m going to buy it.’”
Woo was the first customer to buy an exoskeleton to bring home, paying US $80,000 out of pocket. His insurance wouldn’t cover the cost, but he was able to make the purchase in part because of a legal settlement after his accident. The home-use model came with a requirement that the user have at least one companion who was fully trained in operating the device. In Woo’s case, that meant that Springer learned to suit him up, realign his balance, and help him if he fell.
On delivery day, two SUVs drove up to a hotel down the street from Woo’s condo in the Toronto area. The technicians hauled two huge boxes into a hotel room and assembled his personal exoskeleton. They took Woo’s measurements, made adjustments, checked the software. This latest version could be controlled by either weight shifting or tapping commands on a smartwatch, and Woo had the app ready. He tested out everything in the hotel room, signed off, and then the technicians drove his robot legs to his home.
That was the start of his golden period with the ReWalk—similar to the excitement many people experience with a new piece of exercise equipment. “I used it every day for a few hours, and then I started logging how many steps I’d done,” Woo says. “My last count was probably just slightly over a million steps,” he says, with half of those steps taken in his home unit and half in training programs and clinical trials.
Advertisement
The ReWalk was the first exoskeleton available for use outside the clinic. Robert Woo’s ReWalk arrived in two large boxes. ReWalk engineers assembled it in a hotel room, and Woo tried it out in the hallway before taking it home. Robert Woo
Tristan, Woo’s eldest son, remembers doing laps with his dad in the condo’s underground parking garage while his dad was training for a 5-kilometer race in New York City. Tristan admits that he had previously been embarrassed about his dad, but training for the race shifted something for him. “I was so used to not wanting to tell people that my dad was in a wheelchair, but then I shared his passion for the training,” he says. “When people would come up to us, I’d tell them about it.”
The ReWalk could turn ordinary moments into small engineering projects. On weekends, Woo would take his boys to the golf course behind their condo and bring a baseball. He had rigged two holsters to the sides of the suit so he could stash a crutch and stand on three points (two legs and one arm) while he pitched or caught. Throw, switch crutches, catch. On the day of his accident, he never thought such a scene would be possible. But with the exoskeleton, it became just another design problem to solve. “It’s a little more work. It’s not perfect,” he says. “But in the end, you still get to do what you want to do—which is play ball with your sons.”
Tristan, now a college student, says he didn’t realize at the time how hard his dad worked to make those mundane activities possible. “Reflecting on it now,” he says, “he has shaped almost every element of my life, and he definitely is my hero.”
But even during that golden stretch, the ReWalk had a way of asserting its limits. Every so often it would freeze mid-stride and require a reboot—a small technical hiccup in theory, but a serious problem when there’s a person strapped inside. Once, when he was walking on his own in the parking garage (without his mandated companion), the suit glitched and went into “graceful collapse” mode, lowering him to a seated position on the ground. Woo had to ask security to bring his wheelchair and a dolly.
Advertisement
He had imagined the exoskeleton would be most useful in the kitchen. Woo loves to cook, and he had pictured himself standing at the stove, looking down into pots, and moving easily between counter and sink. The reality, he found out, was more complicated. “It’s actually very time-consuming and troublesome” to cook in an exoskeleton, he says.
Preparing a meal meant first rolling through the kitchen in his wheelchair to gather every ingredient and utensil, then transferring himself into the ReWalk and moving himself into position at the counter, stopping at just the right moment. “That’s when I fell once,” Woo says. “I collided with the counter and then lost my balance and fell backward.” If all went well, he’d lean either on one crutch or the counter to keep his balance while he worked. But if he’d forgotten to grab the vinegar from the cabinet, he’d have to go into walk mode, crutch over to it, and figure out how to carry the bottle back to his workstation.
Sitting unused in Robert Woo’s home, his ReWalk exoskeleton reflects both the promise and the limits of early devices. Robert Woo
Gradually, he stopped trying. The suit, which he’d once worn every day, spent more time sitting idle in the hallway; like so many abandoned treadmills and stationary bikes, it gathered dust. Part of the reason was the exoskeleton’s practical limitations, but part of it was a shocking development: In 2024, Vivian was diagnosed with an aggressive form of breast cancer. She died in November of that year, at the age of 54.
Woo was scheduled to begin a new round of clinical trials for the Wandercraft home-use exoskeleton that month. In the aftermath of Vivian’s death, he postponed his sessions and questioned whether he would ever go back. “At the time, I thought, ‘What’s the point?’” he remembers.
Advertisement
He did go back, though. “He just rolled up, right into my office,” says Mount Sinai’s Riccobono. “He still had Vivian’s box of ashes on his lap. That’s how fresh it was.” Woo brought the box into a meeting of spinal cord injury patients and shared the story of losing the love of his life. And he told them that he heard his wife’s voice in his head every day, telling him to get back to work. Once again, he was figuring out how to move forward with what he had.
How Close Are We to Everyday Exoskeletons?
In the Wandercraft showroom last May, Woo steered toward the door to the street, technicians flanking him like spotters. The slope down to the sidewalk was barely an inch high, but everyone tensed. He shifted his weight and took a step forward. The suit halted automatically. He tried again—step, stop; step, stop—as the suit kept detecting the slight decline and a safety feature kicked in. The Wandercraft isn’t yet rated for slopes of more than 2 percent, and even the gentle pitch of Park Avenue was enough to trigger its safeguards. When he finally reached the sidewalk, Woo broke into a grin. A man in the back seat of a stopped Uber leaned out his window, filming.
During testing of the Wandercraft exoskeleton, straps caused an abrasion on Robert Woo’s leg, which he documented as part of his feedback to the company. Robert Woo
Woo had recently completed seven sessions with the Wandercraft at the VA hospital and had been impressed overall. But at the showroom, he rolled up his pants leg to reveal an abrasion on his shin, the result of a strap that had worn away a patch of skin during a long walking session. He would later send Wandercraft a nine-page assessment with photos and a technology wish list, asking the company to work on things like padding, variable walking speeds, and deeper squats.
Wandercraft’s engineers relish that kind of user feedback, says CEO Matthieu Masselin. Exoskeletons are a far more difficult engineering problem than humanoid robots, he explains. “You basically have two systems of equal importance. You know about the robot—it’s fully quantified and measured. But you don’t know what the person is doing, and how the person is moving within the device.”
Advertisement
Since Woo began testing exoskeletons 15 years ago, both the technology and the market have made strides. ReWalk and Ekso won FDA clearance for clinical use in the 2010s, and both now sell home-use versions. The companies have sold thousands of exoskeletons to rehab clinics and personal users, and they see room for growth; in the United States alone, about 300,000 people live with spinal cord injuries, and millions more have mobility impairments from stroke, multiple sclerosis, or other conditions. The VA began supplying devices to eligible veterans in 2015, and Medicare recently established a system for reimbursement, a move that private insurers are beginning to follow. What was once experimental is slowly becoming established.
Researchers who test the devices say the technology still has significant limits. Pal, of the New Jersey Institute of Technology, mentions battery life, dexterity, and reliability as ongoing challenges. But, he says with a laugh, “Our bodies have evolved over many millions of years—these machines will need a bit more time.” Pal hopes the companies will keep pushing the technological frontier. “My lifetime goal is to see the day when someone like Robert Woo can wake up in the morning, put this device on, and then live an ordinary life.”
For Woo, the real question about the self-balancing Wandercraft was: Could he cook with it? In the VA hospital’s home mockup, he tried it out in the kitchen, stepping sideways to retrieve items from cabinets and squatting to grab something from the fridge’s lower shelf. For the first time in years, he could work at a counter without leaning on crutches. “The self-standing exoskeleton changes everything,” he says. He imagines a user placing a Thanksgiving turkey on a tray attached to the suit and walking it into the dining room.
Back in the showroom, Woo finishes the demo and brings the suit to a seated position before transferring back to his wheelchair. After so many years of testing prototypes, he’s now realistic about the technology’s timeline. A truly all-day exoskeleton—the kind you live in, the kind that replaces a wheelchair—may be a decade or more away. “It may not be for me,” he says. But that’s no longer the point. He’s thinking about young people who are newly injured, who are lying in hospital beds and trying to imagine how their lives can continue. “This will give them hope.”
Audio-Technica has owned a large chunk of the entry-level phono cartridge conversation for years, and the reason is not complicated: its VM95 Series cartridges are affordable, easy to mount, widely supported, and found on a lot of turntables that people can actually afford.
Alongside Ortofon, the Japanese cartridge maker has become one of the default installs on tables below $450, where every dollar matters and cartridge upgrades need to be simple, reliable, and sonically worthwhile.
Now Audio-Technica is expanding that formula with the AT-VM95EBK Dual Moving Magnet Cartridge and AT-VM95EBK/H Headshell/Cartridge Combo Kit, two new black-finished versions built around the same VM95 Series platform.
The cartridge uses a 0.3 x 0.7 mil elliptical stylus, delivers 4.0 mV output, fits standard half-inch mount turntables, and remains compatible with all six interchangeable AT-VMN95 replacement styli.
Advertisement
The cartridge sells for $74, while the pre-mounted headshell combo kit comes in at $109, making this less of a reinvention and more of a smart cleanup job for one of vinyl’s most practical upgrade paths.
Why the VM95 Series Matters
The VM95 Series is one of the reasons Audio-Technica has become such a force in affordable vinyl playback. The concept is simple but effective: one cartridge body, multiple stylus options, broad turntable compatibility, and pricing that does not require a financial intervention from the rest of the household — think about all of the records one can buy that they will never know about if they think you showed some fiscal restraint and stayed below $300.
At the core of the VM95 platform is Audio-Technica’s Vertical Dual Magnet design, which mirrors the 90-degree V-shaped configuration of the cutter head used to create the original vinyl master. Audio-Technica says this helps the cartridge deliver accurate tracking, strong channel separation, a more defined stereo image, and clarity across the frequency range.
The bigger selling point for real-world users is flexibility. Every VM95 cartridge uses the same body design, which means owners can upgrade or replace the stylus without replacing the entire cartridge. The series supports multiple stylus profiles, including conical, elliptical, nude elliptical, Microlinear, Shibata, and 78 RPM conical options. That gives listeners a clear path from an entry-level setup to something more refined without starting over.
Installation is also part of the appeal. All AT-VM95 cartridges fit standard 1/2-inch mount headshells, and the threaded cartridge body allows mounting with two screws and no tiny nuts to drop into the carpet, where they immediately join the witness protection program.
Advertisement
That matters because the VM95 Series is aimed squarely at the part of the market where most vinyl listeners actually live: affordable turntables, modest systems, and users who want better tracking and detail without turning a cartridge upgrade into a weekend engineering project. The new AT-VM95EBK and AT-VM95EBK/H do not change the formula. They make one of Audio-Technica’s most practical cartridge platforms look cleaner in black while keeping the upgrade path intact.
Want More? The AT33x Series Is the Next Step Up
For listeners who want to move beyond the VM95 Series, Audio-Technica’s AT33x Series is the next serious step. Unlike the affordable VM95 moving magnet platform, the AT33x models are moving coil cartridges, handcrafted in Japan and aimed at listeners with better tonearms, more capable phono stages, and records clean enough to tell the truth.
Advertisement. Scroll to continue reading.
The lineup includes three stereo models — AT33xEN, AT33xMLD, and AT33xMLB — plus two mono versions, the AT33xMONO/I and AT33xMONO/II. Prices start at $449 for the mono models and $699 for the stereo versions, topping out at $899 for the AT33xMLB. The range adds more advanced materials, including a die-cast zinc base, hybrid body construction, refined suspension, PCOCC copper coil wiring, and upgraded cantilever/stylus options.
This is where Audio-Technica starts asking more from your system, your setup skills, and your phono stage. Cheap turntable with a built-in phono preamp? Wrong neighborhood. Better deck, proper MC gain, and a little patience? This is where this type of upgrade would make sense. Just don’t tell the family.
Advertisement
The Bottom Line
The Audio-Technica AT-VM95EBK is not a radical new cartridge platform, and that is the point. It brings the proven VM95 Series formula into a cleaner black finish with easy installation, an elliptical stylus, interchangeable stylus upgrades, and strong entry level performance for under $100. The AT-VM95EBK/H combo kit makes even more sense for listeners who want a premounted, ready to install option without turning a simple cartridge upgrade into a lost weekend.
For affordable turntables, this is exactly where Audio-Technica continues to win: practical, upgradeable, widely compatible, and priced for people who still need money left over for records.
NASA has successfully tested an improved flight system designed for Mars’ hostile environment. The new technology can be accelerated beyond the speed of sound (Mach 1), the space agency said, and is expected to significantly enhance the operational capabilities of future exploration missions on the Red Planet. Read Entire Article Source link
Both privilege escalation vulnerabilities stem from bugs in the kernel’s handling of page caches stored in memory, allowing untrusted users to modify them. They target caches in networking and memory-fragment handling components. Specifically, CVE-2026-43284 attacks the esp4 and esp6 () processes, and CVE-2026-43500 zeroes in on rxrpc. Last week’s CopyFail exploited faulty page caching in the authencesn AEAD template process, which is used for IPsec extended sequence numbers. A 2022 vulnerability named Dirty Pipe also stemmed from flaws that allow attackers to overwrite page caches.
Dirty Frag belongs to the same bug family as Dirty Pipe and Copy Fail, but it targets the frag member of the kernel’s struct sk_buff rather than pipe_buffer. The exploit uses splice() to plant a reference to a read-only page-cache page (for example, /etc/passwd or /usr/bin/su) into the frag slot of a sender-side skb. Receiver-side kernel code then performs in-place cryptographic operations on that frag, modifying the page cache in RAM. Every subsequent read of the file sees the corrupted version, even though the attacker only ever had read access.
CVE-2026-43284 is found in the esp_input() process on the IPsec ESP receive path. When an skb object is non-linear but lacks a frag list, the code skips skb_cow_data() and decrypts AEAD in place on the planted frag. From there, an attacker can control the file offset and the 4-byte value of each store.
CVE-2026-43500, meanwhile, resides in rxkad_verify_packet_1(). The process decrypts RxRPC payloads using a single-block process. Splice-pinned pages become both a source and destination. That, paired with the decryption key being freely extracted using the add_key (rxrpc), allows an attacker to rewrite contents in memory.
Advertisement
Either exploit used separately is unreliable. Some Ubuntu configurations use AppArmor to prevent untrusted users from creating namespace contents. That, in turn, neutralizes the ESP technique. Most other distributions by default don’t run rxrpc.ko, which neutralizes the RxRPC arm. When chained together, however, the two exploits allow attackers to obtain root on every major distribution Kim tested. Once the exploits run, attackers can use SSH access, web-shell execution, container escapes, or compromise low-privilege accounts.
“Dirty Frag is notable because it introduces multiple kernel attack paths involving rxrpc and esp/xfrm networking components to improve exploitation reliability,” Microsoft researchers wrote. “Rather than relying on narrow timing windows or unstable corruption conditions often associated with Linux local privilege escalation exploits, Dirty Frag appears designed to increase consistency across vulnerable environments.”
Researchers at Google-owned Wiz said exploits will be less likely to break out of hardened containerized environments such as Kubernets with default security settings in place. “However, the risk remains significant for virtual machines or less restricted environments.”
The best response for anyone using Linux is to install patches immediately. While fixes likely require a reboot, protection from a threat as severe as Dirty Frag outweighs the cost of disruptions. Anyone who can’t install immediately should follow the mitigation steps laid out in the posts linked above. Additional guidance can be found here.
The Entertainment Software Association (ESA) has come out against California bill AB 1921, a state bill that would compel developers to offer remedies before deactivating servers for online games. Stop Killing Games has been fighting this battle for the last couple of years and was quick to condemn the ESA’s position. Read Entire Article Source link
On the lower west side of Manhattan Island, in the Chelsea district, there is an unassuming, concrete-looking townhouse whose previous owners include Lady Gaga and basketball player Kevin Durant. If you ever get the opportunity to saunter through its doors, you’re stepping into a tower of sound.
That’s because the House of Sound, operated by Bose after its acquisition of the Sonus Faber brand, is an ode to audiophile and luxury tastes.
Through six floors of the townhouse, there’s a cadre of McIntosh and Sonus Faber kit, with each room designed to give a taste of what it’d be like to have this hi-fi equipment in your home. As the House of Sound website puts it, it’s a “destination where audio, art, and design intersect to create a truly immersive experience”.
It wants to promote the idea that audio can be considered in the same aspirational league as travel, watches, cars, and haute couture fashion. And that it can be a form of creativity as well, whether that’s through the form it takes – the materials and aesthetic that goes into Sonus Faber and McIntosh products – or how it brings other creative works to life, whether that’s through two-channel stereo or a private cinema install (that Questlove from The Roots rents for the Oscars). It’s the Met Gala for sound systems.
Image Credit (Trusted Reviews)
It’s a place that desires to be the apex of what hi-fi can be – without limitation. It’s not so much a consumable thing that, while enjoyable as an experience, is in ways designed to be disposable. Without trying to sound like an advertisement, you go to its listening rooms to luxuriate in high-fidelity sound, or as it was put to us on the tour, to connect “yourself and your own emotions and the people around you”. Lofty, but why not reach for the stars?
I’ve been to hi-fi demo spaces before, such as KJ West One, which is literally wall-to-wall of high-end hi-fi equipment, or ventured to hi-fi shows such as High End. But the House of Sound obviously feels different from either of those two because it takes place in an actual home space.
If you need to be convinced of parting with money into the six or seven figures, it certainly helps having an idea of how it would look in your own well-appointed home. It all helps to add to the sense of immersion because the space you’re listening in is a familiar-ish one.
Advertisement
Advertisement
Image Credit (Trusted Reviews)
Hi-fi with no limits
Of course, this would be rather moot if the products didn’t sound great. I’ve limited experience with Sonus Faber products, having tested the Omnia all-in-one system several years ago, but Sonus Faber doesn’t really deal with products that tend to be easily shippable.
I’ve heard Amati Supreme hi-fi loudspeakers at events such as 2025’s Paris AV Show, and thought they sounded “phenomenal”. This time I got to hear the Suprema, which is essentially Sonus Faber’s no-holds-barred loudspeaker.
Image Credit (Trusted Reviews)
And they sounded phenomenal. They’re a bit on the crisp side of neutral, so at times can sound a little thin to my ears, but they generate huge levels of transparency, insight and naturalism, as well as power and energy in a stereo image that’s wide and deep, with minimal, if any, distortion.
We got to play a selection of tracks*, to just sit there and listen to the speakers. Some people didn’t want to leave.
Advertisement
Advertisement
*(If you want to know my choices, they were Slipknot’s Duality and Illit’s K-pop Magnetic in an attempt to try and ‘break’ the speakers. I failed.)
My own private cinema
We then descended to the ground floor (first floor for any Americans reading) and had the opportunity to listen to the private home cinema install.
Past a large, nondescript door that doesn’t hint at the excitement that awaits, is a reference standard private home cinema with a sound system that will blow your Sonos Arc Ultra surround system away.
Dotted around the room is a 29-channel system that includes Sonus Faber Arena 20 in-wall speakers, Arena 10 in-ceiling modules, and Arena 30 speakers behind the screen in a left, centre, right configuration with a dual-tweeter design similar to the Amati Supreme for clearer dialogue. In total, there are 16 (sixteen!) subwoofers.
Advertisement
Image Credit (Trusted Reviews)
Advertisement
It’s powered by 19 McIntosh amplifiers that, apparently, provide a whopping 22,400W of total power to the system. All amps have a THD of less than 0.005% for absolutely minimal distortion, and the amps power on in a trigger sequence to avoid a massive on-rush of current when you’ve got 20,000+ watts waiting to be released.
An interesting little titbit was the reveal that in films, there’s generally only 8-9 minutes of true LFE (Low Frequency Effects) in a typical two-hour film. 12 of the sixteen subwoofers are then repurposed with the other channels, with the front left/right receiving a dedicated cluster of subs, and the sides, rears, and even the ceiling arrays partnered with a sub.
The result of this configuration was full-range sound from infrasonic to beyond audible high frequency, allowing for precise placement of bass in an area of the room rather than just shaking the entire floor.
This private cinema is placed on the first floor under the kitchen, and apparently, you can feel the rumble in the kitchen even if you can’t hear what’s playing. That’s the power unleashed by this cinema.
Advertisement
Image Credit (Trusted Reviews)
Kaleidescape is the source for films, with an Apple TV nearby for sports and streaming, plus a PS5 for gaming.
Advertisement
And we were treated to Top Gun: Maverick, which has become a staple of Dolby Atmos demos (everyone from Sonos, Yamaha and Focal uses it, moving on from Mad Max: Fury Road being).
It is probably (memory aside), the best I’ve heard the film since watching it in Dolby Cinema at the West End Odeon (the better of the Odeon Leicester Square cinemas). It sounded immense, the nuance of the smaller details that might be lost in a home cinema set-up are rendered crystal clear. The system has even been given the thumbs up by the Oscar-winning sound mixer of Maverick, who watched the film there at an event.
The best home cinema systems can put you in an immersive bubble, whether it’s object-based or channel-based. The private cinema in this townhouse is an experience where you feel it too… and you don’t have to bother with people talking or a crisp packet rustling in the darkness, as it did when I watched The Drama a few weeks ago and annoyed another patron in the cinema.
Advertisement
Head out on the highway
So I’ve written about hi-fi and home cinema. Why not cars too?
On the same first floor as the private cinema is a Lamborghini tucked away in the corner. Inside is a Sonus Faber sound system that’s been tuned for the (tight) interior environment. I’ve written in the past how, for many people, a car might be the best way to listen to music, and it’s the same case for this Lamborghini system.
The system itself is not as numerous in speakers or has quite as fancy custom technology as the Bowers & Wilkins kit in the Polestar 3, but the sense of immersion convinces me that cars make for a pretty excellent hi-fi room. The low end produced, despite there being no dedicated sub (if memory serves), brought genuine bass to the proceedings, but the best thing about the whole experience is how balanced it all sounded.
Advertisement
Advertisement
I’m still unnerved by my own anxiety that bass would distract during a car trip, but then wouldn’t the roar of the engine distract too? Perhaps they’d cancel each other out.
Luxury, aspirational sound
Image Credit (Trusted Reviews)
It was a great few hours at the House of Sound alongside seeing Bose’s Lifestyle Collection, products which aim for premium but for a mainstream audience. The House of Sound shows the potential for hi-fi to move into more luxurious realms (if it hasn’t already).
Of course, this is not an area that I or most people who happen across this article would ever find themselves inhabiting. The Aida sound system, along with all the McIntosh equipment in the room, is an easy seven-figure cost. These are sound systems that would exist in people’s dreams.
But for a few hours in Bose’s House of Sound, those dreams can become reality.
Nvidia’s real AI moat isn’t “a piece of hardware,” writes Wired’s Sheon Han. It’s CUDA: a mature, deeply optimized software ecosystem that keeps machine-learning workloads tied to Nvidia GPUs. An anonymous reader quotes a report from Wired: What sounds like a chemical compound banned by the FDA may be the one true moat in AI. CUDA technically stands for Compute Unified Device Architecture, but much like laser or scuba, no one bothers to expand the acronym; we just say “KOO-duh.” So what is this all-important treasure good for? If forced to give a one-word answer: parallelization. Here’s a simple example. Let’s say we task a machine with filling out a 9×9 multiplication table. Using a computer with a single core, all 81 operations are executed dutifully one by one. But a GPU with nine cores can assign tasks so that each core takes a different column — one from 1×1 to 1×9, another from 2×1 to 2×9, and so on — for a ninefold speed gain. Modern GPUs can be even cleverer. For example, if programmed to recognize commutativity — 7×9 = 9×7 — they can avoid duplicate work, reducing 81 operations to 45, nearly halving the workload. When a single training run costs a hundred million dollars, every optimization counts.
Nvidia’s GPUs were originally built to render graphics for video games. In the early 2000s, a Stanford PhD student named Ian Buck, who first got into GPUs as a gamer, realized their architecture could be repurposed for general high-performance computing. He created a programming language called Brook, was hired by Nvidia, and, with John Nickolls, led the development of CUDA. If AI ushers in the age of a permanent white-collar underclass and autonomous weapons, just know that it would all be because someone somewhere playing Doom thought a demon’s scrotum should jiggle at 60 frames per second. CUDA is not a programming language in itself but a “platform.” I use that weasel word because, not unlike how The New York Times is a newspaper that’s also a gaming company, CUDA has, over the years, become a nested bundle of software libraries for AI. Each function shaves nanoseconds off single mathematical operations — added up, they make GPUs, in industry parlance, go brrr.
A modern graphics card is not just a circuit board crammed with chips and memory and fans. It’s an elaborate confection of cache hierarchies and specialized units called “tensor cores” and “streaming multiprocessors.” In that sense, what chip companies sell is like a professional kitchen, and more cores are akin to more grilling stations. But even a kitchen with 30 grilling stations won’t run any faster without a capable head chef deftly assigning tasks — as CUDA does for GPU cores. To extend the metaphor, hand-tuned CUDA libraries optimized for one matrix operation are the equivalent of kitchen tools designed for a single job and nothing more — a cherry pitter, a shrimp deveiner — which are indulgences for home cooks but not if you have 10,000 shrimp guts to yank out. Which brings us back to DeepSeek. Its engineers went below this already deep layer of abstraction to work directly in PTX, a kind of assembly language for Nvidia GPUs. Let’s say the task is peeling garlic. An unoptimized GPU would go: “Peel the skin with your fingernails.” CUDA can instruct: “Smash the clove with the flat of a knife.” PTX lets you dictate every sub-instruction: “Lift the blade 2.35 inches above the cutting board, make it parallel to the clove’s equator, and strike downward with your palm at a force of 36.2 newtons.” “You can begin to see why CUDA is so valuable to Nvidia — and so hard for anyone else to touch,” writes Han. “Tuning GPU performance is a gnarly problem. You can’t just conscript some tender-footed undergrad on Market Street, hand them a Claude Max plan, and expect them to hack GPU kernels. Writing at this level is a grindsome enterprise — unless you’re a cracker-jack programmer at DeepSeek…”
Han goes on to argue that rivals like AMD and Intel offer competitive specs on paper, but their software stacks have struggled with bugs, compatibility issues, and weak adoption. As a result, Nvidia has built an Apple-like moat around AI computing, leaving the industry dependent on its expensive hardware.
Climbing into the open metal cage of Unitree’s GD01 feels like slipping behind the wheel of something from another era of imagination. Founder Wang Xingxing does exactly that in the one-minute demonstration video released today. He buckles into the central seat, grips the controls, and sets the machine in motion across an indoor workshop floor. The robot responds with smooth, deliberate steps on its beastly red legs.
From the outside, you can tell how powerful this beast is; shiny red panels cover the limbs and torso frame, while silver bars keep everyone secure within the open cockpit. It’s imposing, easily twice the height of most adults, and you get a terrific view from up top, but you also get a sense of how large this thing is. There are thick black treads wrapping around the frame and feet in case it needs additional traction, but then there are hydraulic lines and joint housings running down the arms and legs, which gives you a real sense of what’s propelling this beast.
Sleek & Durable Design: Standing at 132cm tall and weighing only approx. 35kg, the G1 is constructed with aerospace-grade aluminum alloy and carbon…
High Flexibility & Safe Movement: Boasting 23 joint degrees of freedom (6 per leg, 5 per arm), it offers an extensive range of motion. For safety, it…
Smart Interaction & Connectivity: Powered by an 8-core high-performance CPU and equipped with a depth camera and 3D LiDAR. It supports Wi-Fi 6 and…
Power shows up clearly when Wang guides the bipedal form toward a stack of bricks. One solid push from the body and the pile collapses. No extra tools or dramatic wind-up needed. The 500-kilogram total weight, including the pilot, delivers real structural strength without any loss of control. Unitree notes the machine works as a civilian vehicle built for practical jobs like transport across rough sites, basic exploration, or even rescue work where a tall vantage point helps. Pricing starts at 3.9 million yuan, which works out to roughly 650 thousand dollars. That figure covers the base model now headed into mass production. Buyers will get a complete, ready-to-pilot system rather than a kit or prototype.
Sam Altman greets Microsoft CEO Satya Nadella at OpenAI DevDay in San Francisco in 2023. (GeekWire File Photo / Todd Bishop)
Satya Nadella drew a historical parallel to Microsoft’s early PC partnership with IBM as the tech giant prepared to invest $10 billion more in OpenAI in April 2022 — writing in an internal email that he didn’t want Microsoft to become IBM while OpenAI became the next Microsoft.
That email, presented as evidence by Elon Musk’s lead trial attorney Steven Molo, was one of the new details to emerge from the Microsoft CEO’s turn on the stand Monday morning in Musk’s lawsuit against Sam Altman, OpenAI and Microsoft in federal court in Oakland.
Nadella described the decision to invest in OpenAI as a “one-way door,” saying Microsoft couldn’t build two supercomputers — one for itself and one for OpenAI — and had to accept the opportunity cost of diverting scarce computing resources away from its own AI teams.
“We were outsourcing essentially a lot of the core IP development and taking a massive dependency on OpenAI,” Nadella testified, explaining that he wanted to ensure Microsoft had access to the intellectual property generated by the partnership, and continued to build its own knowledge and capabilities at the same time.
Board considerations unredacted: The testimony also provided new information from messages among Microsoft execs and Altman in the days following his brief ouster as OpenAI CEO in 2023. The names of potential candidates from that thread were previously redacted in public court records.
Advertisement
From Nadella’s testimony Monday, it emerged that two potential OpenAI board candidates for whom he voiced his disapproval were Diane Greene, the former Google Cloud CEO, and Bing Gordon, the veteran gaming exec and Kleiner Perkins partner previously on Amazon’s board. Nadella said he objected to both as potential candidates because of their ties to companies that compete directly with Microsoft in AI.
He said the discussions were initiated by Altman and other OpenAI insiders seeking his input, and that the board could have ignored his suggestions. One candidate he suggested, former Gates Foundation CEO Sue Desmond-Hellman, was later appointed to the board.
Musk argues that Microsoft’s efforts to protect its interests in the OpenAI partnership came at the expense of the OpenAI nonprofit’s original mission to develop AI for the benefit of humanity. His lawsuit alleges that Microsoft aided and abetted a breach of the charitable trust that governed OpenAI’s founding, misusing his original investment, estimated at $38 million to $44 million.
Enabling a massive nonprofit: Nadella offered a different view on the stand, describing a collaboration built on mutual benefit in which Microsoft took on enormous risk to support a fledgling AI lab that no one else was willing to fund. He said the partnership had created “one of the largest nonprofits in the world,” enabling products like ChatGPT and Copilot that put AI tools in the hands of millions of people.
Advertisement
Under cross-examination, however, Nadella acknowledged that he was not aware of any full-time employees at the OpenAI nonprofit before March 2026, or of any grants, research, or open-sourced technology it had produced.
One of Microsoft’s attorneys in the case, Jay Jurata of Dechert, also sought to undermine Musk’s standing in the case. He walked Nadella through three major milestones in the Microsoft-OpenAI partnership — the 2019 announcement, a 2020 exclusive license to GPT-3, and the 2023 $10 billion investment — and asked each time whether Musk had reached out to object.
Each time, Nadella said no. He and Musk have each other’s phone numbers, he added.
Microsoft estimates the OpenAI return: Musk’s attorney, on cross-examination, sought to show the benefits Microsoft has received from the partnership. He walked Nadella through a January 2023 memo from Microsoft President Brad Smith to the company’s board, projecting a $92 billion return on Microsoft’s cumulative $13 billion investment in OpenAI.
Advertisement
According to the testimony, a footnote in the memo showed a 20% annual increase kicking in starting in 2025, which could roughly double the return within four years.
Under the restructured deal announced last year, the caps on Microsoft’s returns were removed entirely. Microsoft and OpenAI also recently amended the partnership to make Microsoft’s IP license non-exclusive and open all OpenAI products to any cloud provider.
[Update: The Informationreported Monday that revenue-sharing payments from OpenAI to Microsoft under the new deal are capped at $38 billion.]
Asked about the memo on the witness stand, Nadella confirmed the figures but noted that the investment carried real risk, saying the return could just as easily have been zero.
Advertisement
The trial, before U.S. District Judge Yvonne Gonzalez Rogers, is expected to continue through May 21, with OpenAI CEO Sam Altman also expected to take the stand this week.
GeekWire reported on today’s proceedings via the court’s audio livestream. Correction: The name of Microsoft’s outside counsel for Nadella’s testimony has been corrected since publication.
A doctor in a hospital exam room watches as a medical transcription agent updates electronic health records, prompts prescription options, and surfaces patient history in real time. A computer vision agent on a manufacturing line is running quality control at speeds no human inspector can match. Both generate non-human identities that most enterprises cannot inventory, scope, or revoke at machine speed.
That is the structural problem keeping agentic AI stuck in pilots. Not model capability. Not compute. Identity governance.
Cisco President Jeetu Patel told VentureBeat at RSAC 2026 that 85% of enterprises are running agent pilots while only 5% have reached production. That 80-point gap is a trust problem. The first questions any CISO will ask: which agents have production access to sensitive systems, and who is accountable when one acts outside its scope? IANS Research found that most businesses still lack role-based access control mature enough for today’s human identities, and agents will make it significantly harder. The 2026 IBM X-Force Threat Intelligence Index reported a 44% increase in attacks exploiting public-facing applications, driven by missing authentication controls and AI-enabled vulnerability discovery.
Why the trust gap is architectural, not just a tooling problem
Michael Dickman, SVP and GM of Cisco’s Campus Networking business, laid out a trust framework in an exclusive interview with VentureBeat that security and networking leaders rarely hear stated this plainly. Before Cisco, Dickman served as Chief Product Officer at Gigamon and SVP of Product Management at Aruba Networks.
Advertisement
Dickman said that the network sees what other telemetry sources miss: actual system-to-system communications rather than inferred activity. “It’s that difference of knowing versus guessing,” he said. “What the network can see are actual data communications … not, I think this system needs to talk to that system, but which systems are actually talking together.” That raw behavioral data, he added, becomes the foundation for cross-domain correlation, and without it, organizations have no reliable way to enforce agent policy at what he called “machine speed.”
The trust prerequisite that most AI strategies skip
Dickman argues that agentic AI breaks a pattern he says defined every prior technology transition: deploy for productivity first, bolt on security later.
“I don’t think trust is one of those things where the business productivity comes first, and the security is an afterthought,” Dickman told VentureBeat. “Trust actually is one of the key requirements. Just table stakes from the beginning.”
Observing data and recommending decisions carries consequences that stay contained. Execution changes everything. When agents autonomously update patient records, adjust network configurations, or process financial transactions, the blast radius of a compromised identity expands dramatically.
Advertisement
“Now more than ever, it’s that question of who has the right to do what,” Dickman said. “The who is now much more complicated because you have the potential in our reality of these autonomous agents.”
Dickman breaks the trust problem into four conditions. The first is secure delegation, which starts by defining what an agent is permitted to do and maintaining a clear chain of human accountability. The second is cultural readiness; he pointed to alert fatigue as a case study. The traditional fix, Dickman noted, was to aggregate alerts, so analysts see fewer items. With agents capable of evaluating every alert, that logic changes entirely.
“It is now possible for an agent to go through all alerts,” Dickman said. “You can actually start to think about different workflows in a different way. And then how does that affect the culture of the work, which is amazing.”
The third is token economics: Every agent’s action carries a real computational cost. Dickman sees hybrid architectures as the answer, where agentic AI handles reasoning while traditional deterministic tools execute actions. The fourth is human judgment. For example, his team used an AI tool to draft a product requirements document. The agent produced 60 pages of repetitive filler that immediately provided how technically responsive the architecture was, yet showed signs of needing extensive fine-tuning to make the output relevant. “There’s no substitute for the human judgment and the talent that’s needed to be dextrous with AI,” he said.
Advertisement
What the network sees that endpoints miss
Most enterprise data today is proprietary, internal, and fragmented across observability tools, application platforms, and security stacks. Each domain team builds its own view. None sees the full picture.
“It’s that difference of knowing versus guessing,” Dickman said. “What the network can see are actual data communications. Not ‘I think this system needs to talk to that system,’ but which systems are actually talking together.”
That telemetry grows more valuable as IoT and physical AI proliferate. Computer vision agents analyzing shopper behavior and running factory-floor quality control generate highly sensitive data that demands precise access controls.
“All of those things require that trust that we started with, because this is highly sensitive data around like who’s doing what in the shop or what’s happening on the factory floor,” Dickman said.
Advertisement
Why siloed agent data misses the signal
“It’s not only aggregation, but actually the creation of knowledge from the network,” Dickman said. “There are these new insights you can get when you see the real data communications. And so now it becomes what do we do first versus second versus third?”
That last question reveals where Dickman’s focus lands: the strategic challenge is sequencing, not capability.
“The real power comes from the cross-domain views. The real power comes from correlation,” Dickman said. “Versus just aggregation and deduplication of alerts, which is good, but it’s a little bit basic.”
This is where he sees the most common pitfall. Team A builds Agent A on top of Data A. Team B builds Agent B on top of Data B. Each silo produces incrementally useful automation. The cross-domain insight never materializes.
Advertisement
Independent practitioners validate the pattern. Kayne McGladrey, an IEEE senior member, told VentureBeat that organizations are defaulting to cloning human user profiles for agents, and permission sprawl starts on day one. Carter Rees, VP of AI at Reputation, identified the structural reason. “A significant vulnerability in enterprise AI is broken access control, where the flat authorization plane of an LLM fails to respect user permissions,” Rees told VentureBeat. Etay Maor, VP of Threat Intelligence at Cato Networks, reached the same conclusion from the adversarial side. “We need an HR view of agents,” Maor told VentureBeat at RSAC 2026. “Onboarding, monitoring, offboarding.”
Agentic AI trust gap assessment
Use this matrix to evaluate any platform or combination of platforms against the five trust gaps Dickman identified. Note that the enforcement approaches in the right column reflect Cisco’s framework.
Trust gap
Current control failure
Advertisement
What network-layer enforcement changes
Recommended action
Agent identity governance
IAM built for human users cannot inventory, scope, or revoke agent identities at machine speed
Advertisement
Agentic IAM registers each agent with defined permissions, an accountable human owner, and a policy-governed access scope
Audit every agent identity in production. Assign a human owner. Define permitted actions before expanding the scope
Blast radius containment
Host-based agents and perimeter controls can be bypassed; flat segments give compromised agents lateral movement
Advertisement
Microsegmentation enforces least-privileged access at the network layer, limiting blast radius independent of host-level controls
Implement microsegmentation for every agent-accessible system. Start with the highest-sensitivity data (PHI, financial records)
Cross-domain visibility
Siloed observability tools create fragmented views; Team A’s agent data never correlates with Team B’s security telemetry
Advertisement
Network telemetry captures actual system-to-system communications, feeding a unified data fabric for cross-domain correlation
Unify network, security, and application telemetry into a shared data fabric before deploying production agents
Governance-to-enforcement pipeline
No formal process connecting business intent to agent policy to network enforcement
Advertisement
Policy-to-enforcement pipeline translates governance decisions into machine-speed network rules
Establish a formal pipeline from business-intent definition to automated network policy enforcement
Cultural and workflow readiness
Organizations automate existing workflows rather than redesigning for agent-scale processing
Advertisement
Network-generated behavioral data reveals actual usage patterns, informing workflow redesign
Run a 30-day telemetry capture before designing agent workflows. Build around observed data, not assumptions
A broken ankle and a microsegmentation lesson
Dickman grounded his framework in a scenario from his own life. A family member recently broke an ankle, which put him in a hospital exam room watching a medical transcription agent update the EHR, prompt prescription options, and surface patient history in real time. The doctor approved each decision, but the agent handled tasks that previously required manual entry across multiple systems.
The security implications hit differently when it is a loved one’s records on the screen.
Advertisement
“I would call it do governance slowly. But do the enforcement and implementation rapidly,” he said. “It must be done in machine speed.”
It starts with agentic IAM, where each agent is registered with defined permitted actions and a human accountable for its behavior.
“Here’s my set of agents that I’ve built. Here are the agents. By the way, here’s a human who’s accountable for those agents,” Dickman said. “So if something goes wrong, there’s a person to talk to.”
That identity layer feeds microsegmentation — a network-enforced boundary Dickman says enforces least-privileged access and limits blast radius.
Advertisement
“Microsegmentation guarantees that least-privileged access,” Dickman said. “You’re not relying on a bunch of host agents, which can be bypassed or have other issues.”
If the governance model works for a medical transcription agent handling patient records in an emergency department, it scales to less sensitive enterprise use cases.
Five priorities before agents reach production
1. Force cross-functional alignment now. Define what the organization expects from agentic AI across line-of-business, IT, and security leadership. Dickman sees the human coordination layer moving more slowly than the technology. That gap is the bottleneck.
2. Get IAM and PAM governance production-ready for agents. Dickman called out identity and access management and privileged access management specifically as not mature enough for agentic workloads today. Solidify the governance before scaling the agents. “That becomes the unlock of trust,” he said. “Because when the technology platform is ready, you then need the right governance and policy on top of that.”
Advertisement
3. Adopt a platform approach to networking infrastructure. A platform strategy enables data sharing across domains in ways fragmented point solutions cannot. That shared foundation is what makes the cross-domain correlation in the trust gap assessment above operationally real.
4. Design hybrid architectures from the start. Agentic AI handles reasoning and planning. Traditional deterministic tools execute the actions. Dickman sees this combination as the answer to token economics: it delivers the intelligence of foundation models with the efficiency and predictability of conventional software. Do not build pure-agent systems when hybrid systems cost less and fail more predictably.
5. Make the first use cases bulletproof on trust. Pick two or three high-value use cases and build them with role-based access control, privileged access management, and microsegmentation from day one. Even modest deployments delivered with best practices intact build the organizational confidence that accelerates everything after.
“You can guarantee that trust to the organization, and that will unleash the speed,” Dickman said.
Advertisement
That is the structural insight running through every section of this conversation. The 85% of enterprises stuck in pilot mode are not waiting for better models. They are waiting for the identity governance, the cross-domain visibility, and the policy enforcement infrastructure that makes production deployment defensible. Whether they build on Cisco’s platform or assemble their own, Dickman’s framework holds: identity governance, cross-domain visibility, policy enforcement. None of those prerequisites is optional.
The organizations that satisfy them first will deploy agents at a pace the rest cannot match, because every new agent inherits the trust architecture the first ones required. The ones still debating whether to start will watch that gap widen. Theoretical trust does not ship.
Meanwhile, Sutskever testified that he had raised concerns about Sam Altman because he feared OpenAI could be “destroyed.” He expressed concerns about Altman’s behavior to the board, in part because he said he felt “a great deal of ownership” over the startup. “I simply cared for it, and I didn’t want it to be destroyed,” Sutskever said. CNBC reports: Nadella said he was “very proud” that Microsoft took the risk to invest in OpenAI when “no one else was willing” to bet on the fledgling lab. Musk, who testified late last month, said Microsoft’s $10 billion investment was the key tipping point that made him believe OpenAI was violating its nonprofit mission. He testified that the scale of the investment bothered him, and it prompted him to open a legal investigation into OpenAI. “I was concerned they were really trying to steal the charity,” Musk said from the stand.
Nadella said he did not believe Microsoft’s investments in OpenAI were donations, and that there was a clear commercial element to their partnership from the outset. He said during the partnership’s early years, Microsoft gave OpenAI sharp discounts on computing resources, and Microsoft believed it would reap marketing benefits from doing so. During a separate video deposition that was played on Monday morning, Michael Wetter, a corporate development executive at Microsoft, said the company has recognized approximately $9.5 billion in revenue to date through its partnership with OpenAI as of March 2025.
[…] Nadella said he was “pretty surprised” by the board’s decision [to fire Altman in November 2023], and that his priority was to try and figure out how to maintain continuity for Microsoft customers. Immediately after Altman was removed, Nadella said he made an effort to learn more about what happened, adding that he suspected jealousy and poor communication was at play. During conversations with OpenAI board members after the firing, Nadella said he was simply trying to understand the language in the OpenAI’s statement about Altman being “not consistently candid” while communicating with the board. That language, Nadella said, “just didn’t sort of suffice, because this is the CEO of a company that we are invested in and we’re deeply partnered with, and so I felt that they could have explained to me what are the incidents or what is the detail behind it.” There must have been instances of jealousy or miscommunication that could have justified pushing out Altman, Nadella said. He wanted more depth from the board members after the remark about candor, but no such information was available, he said. “It was sort of amateur city, as far as I’m concerned,” Nadella testified.
Advertisement
[…] Musk testified that he is not entirely against OpenAI having a for-profit unit, but he said it became “the tail wagging the dog.” He repeatedly accused Altman and Brockman of enriching themselves from a charity while also reaping the positive associations that come from running a nonprofit. “Microsoft has their own motivations, and that would be different from the motivations of the charity,” Musk said from the stand. “All due respect to Microsoft, do you really want Microsoft controlling digital superintelligence?”
During a videotaped deposition shown in court last week, former OpenAI director Tasha McCauley recalled a discussion with Nadella and her fellow board members after the 2023 decision to dismiss Altman as OpenAI’s CEO. “To the best of my recollection, Satya wanted to restore things to as they had been,” McCauley said. The board members didn’t think that was the right move, she said. But as a court witness on Monday, Nadella said he never demanded that the board reinstate Altman as OpenAI CEO. Recap:
You must be logged in to post a comment Login