AI/ML News & Innovations Hub

AI/ML news, top picks, and generated innovation digests.

★ Visit ai-karthik.com
422Sources
60663News Items
8Top Picks
322Blogs
failedLast Run

Robotics

200 articles tagged with this keyword, sorted by most recent first.

← All Keywords
LessWrong AI 2026-09-28 17:26 UTC Score 68.0 USR-0152-20260928-community-fo-b5e996b7

9 reasons against a near-term AI slow-down

I’ve heard more arguments in favor of a near-term [1] coordinated AI pause/slow-down [2] than arguments against. I think such questions, including complex flow-through effects, are difficult, and it’s important to really consider both sides. I am not convinced by these arguments, and feel highly uncertain as to when a slow-down would be best, but believe it most likely will be good to slow down or pause AI development at some point. Here are 10 reasons against a near-term AI slow-down: A near-term slowdown could prevent the warning shots that would make the public take AI risk more seriously. And techniques for safe near-term AI may be insufficient for advanced AI, so a near-term slow-down could give a false sense of security for handling truly dangerous AI. Slowing down too early could have an idea inoculation / boy-calls-wolf effect, where the topic becomes more politicized [3] and taken less seriously because people haven’t see anything scary enough to make them see AI as truly dangerous, and then when we really need to go slow later it’s much more difficult to get political will. Not slowing down means we have less of various types of overhang that could make things more dangerous if/when full-speed research resumes, such as: Compute overhang (if hardware is not also fully paused) and other technical advances that complement AI Progress on robotics More AI researchers joining the field Overhang of new ideas for training AI more effectively Progress in biology that lowers…

Entrackr AI 2026-09-28 16:14 UTC Score 80.0 USR-0212-20260928-regional-new-46ad1057

Physical AI company SiMa.ai raises $150 Mn in Series C round

Physical AI company SiMa.ai has raised $150 million in a Series C financing round, bringing its total capital raised to $500 million and valuing the company at $1.45 billion. The round was co-led by Fidelity Management & Research Company and Amplify, with participation from Alter Venture Partners, Dell Technologies Capital and StepStone Group. AllianceBernstein, Baron Capital and J.P. Morgan also joined the round as new investors. The proceeds will be used to scale Palette Neat, an agentic software environment for Physical AI, and develop next-generation hardware capable of delivering 1,000 TOPS of compute through purpose-built Physical AI silicon, SiMa.ai said in a press release. Founded in 2018 by Krishna Rangasayee, SiMa.ai provides a software-centric platform for Physical AI applications. The company focuses on robotics, automotive, drones, industrial automation, aerospace and defence, smart vision and healthcare. SiMa.ai said it serves more than 150 customers across automotive, drones and robotics, including ARK Electronics, AVerMedia, Bosch, Emerson, Intrinsic, Kontron, L&T Technology Services, Mistral, STIGA, Synopsys and Virya Autonomous Tech, among others. According to market research cited by the company, the global Physical AI devices market, including robotics, automotive and drones, is projected to reach 145 million cumulative shipments by 2035. SiMa.ai said Physical AI applications have traditionally relied on NVIDIA GPUs, which can be expensive and power inten…

The Verge AI 2026-09-28 14:55 UTC Score 57.0 AI-016-20260928-global-ai-ne-11285ad7

Dreame’s step-climbing X50 Ultra mopping vacuum is hundreds off

It’s been a while since we’ve seen a big price cut on a capable robot vacuum and mop hybrid that can scale over short steps and threshold to hunt messes. Dreame’s X50 Ultra is the model that my colleague Jen recommends for hard floors and carpets, and it’s down to $899.99 at Amazon for Prime […]

The Verge AI 2026-09-28 08:00 UTC Score 59.0 AI-016-20260928-global-ai-ne-68e35640

Honor’s Magic 9 Pro Max has a big camera and a bigger battery

Honor launches its new Magic 9 flagship phones in China today, and the 9 Pro Max features a design revamp, a capable camera, and a colossal battery. This is Honor's first flagship launch - Robot Phone aside - since it announced its collaboration with camera company Arri. The initial impact is twofold. The new camera […]

Korea AI Times 2026-09-28 06:45 UTC Score 42.0 USR-0048-20260928-global-ai-ne-c8e0003f

우크라이나, '로봇 군단' 발족...드론 이어 지상 전장 무인화 박차

우크라이나가 지상 전장의 무인화를 본격적으로 추진한다. 2022년 시작한 ‘드론 군단(Army of Drones)’에 이어 이번에는 지상 로봇을 전쟁의 핵심 전력으로 끌어올리는 ‘로봇 군단(Army of Robots)’ 구상을 내놨다.미하일로 페도로프 전 우크라이나 국방부 장관은 26일(현지시간) 소셜미디어 X를 통해 “우리는 로봇 군단을 출범시킨다”며 “이제 필요한 것은 전쟁의 로봇화라는 다음 기술적 돌파구”라고 밝혔다.페도로프 전 장관은 드론 분야에서 구축한 상당한 기반을 바탕으로 지상 로봇 기술 전환을 확대한다고 밝혔다. 현재

Nature Machine Intelligence 2026-09-28 00:00 UTC Score 65.0 AI-025-20260928-global-ai-ne-b559b29f

Minute-scale training for microrobot navigation

Nature Machine Intelligence, Published online: 28 September 2026; doi:10.1038/s42256-026-01305-w A vectorized simulator and structured reward framework enable microrobot navigation policies to be trained within minutes and transferred without retraining across robots and environments.

The Decoder 2026-09-27 10:59 UTC Score 53.0 AI-168-20260927-regional-ai--811e688d

Researchers plug GPT-6 Astra directly into a robot and let it clean up an unfamiliar kitchen

Researchers from Stanford and Caltech had a humanoid robot powered by GPT-6 Astra independently tidy up an unfamiliar kitchen. Their HomeBody system skips a specially trained control layer, letting the language model call directly into modular skills like grasping and navigating. The article Researchers plug GPT-6 Astra directly into a robot and let it clean up an unfamiliar kitchen appeared first on The Decoder .

The Decoder 2026-09-26 19:10 UTC Score 41.0 AI-168-20260926-regional-ai--1750e0d2

Former Ukrainian Defense Minister Fedorov pitches a private-sector robot army

Former Ukrainian Defense Minister Mykhailo Fedorov has announced "Army of Robots," a private combat robotics initiative. The robots would handle casualty evacuation, mine clearance, and combat. Drones already account for 95 percent of target engagements, he says. The article Former Ukrainian Defense Minister Fedorov pitches a private-sector robot army appeared first on The Decoder .

Synced 2026-09-26 12:32 UTC Score 51.0 AI-041-20260926-ai-specialis-94ac4391

Comment on Yann LeCun Team’s New Research: Revolutionizing Visual Navigation with Navigation World Models by Leo

Simulating possible routes before acting shifts the problem from purely reactive control toward planning with an internal model. The crucial issue may be how reliably predicted video outcomes correspond to real-world constraints: a visually plausible path could still conceal obstacles, unsafe terrain, or localization errors. Evaluating feasibility should therefore include uncertainty calibration and recovery behavior, not just whether the generated trajectory looks valid. It would also be useful to compare the computational cost of testing several imagined plans with the benefit gained in unfamiliar or hazardous environments where navigation errors are especially consequential.

iAfrica 2026-09-26 10:31 UTC Score 57.0 AI-151-20260926-regional-ai--dba474ec

Kenyan Startup Builds Robotic Sign Language Interpreter — and the African Dataset to Train It

Kenyan computer scientist Norah Kimathi has built a robotic system that translates a teacher’s spoken words into sign language in real time — and, alongside it, a database of African sign languages to train the models that make it work. Kimathi, 22, co-founded ZeroBionic in 2023 and completed her BSc in Computer Science at Strathmore [...]

South China Morning Post AI 2026-09-26 05:45 UTC Score 41.0 AI-156-20260926-regional-ai--e991308a

China’s AI-led transformation is taking shape. Why are some economists urging caution?

As AI becomes an increasingly important part of China’s economic transformation, economists and industry leaders are raising concerns about two potential risks: inflated valuations for humanoid robotics companies and the technology widening the wealth gap between rich and poor. Daniel Zhang, managing partner of FirstLight Capital and former chairman and CEO of Alibaba Group Holding, said the sharp fall in some Chinese humanoid robotics stocks reflected expectations that had become...

The Verge AI 2026-09-25 17:01 UTC Score 60.0 AI-016-20260925-global-ai-ne-78bf1852

Tesla’s Optimus robot is going through growing pains

Hitting its goal of making 20,000 Optimus robots per week is reportedly proving tricky for Tesla. The Information reports that Tesla produced "several hundred robots a week" last month, after it repurposed its Model S and Model X production lines for Optimus earlier this year. However, this strategy is reportedly creating manufacturing snags, like issues […]

IEEE Spectrum Machine Learning 2026-09-24 18:00 UTC Score 58.0 AI-020-20260924-global-ai-ne-7a8c15db

Mexican EPICS in IEEE Team Builds Portable Educational Platform

In Guadalajara, Mexico, many high schools have motivated teachers and talented students with an interest in science, technology, engineering, and mathematics, but they lack access to advanced tools such as robotics laboratories. The resources shortfall limits the students’ opportunities for hands-on learning on cutting-edge applications. A team from ITESO, Universidad Jesuita de Guadalajara , is working to change that. Through the EPICS in IEEE initiative, a multidisciplinary group of 15 engineering students, faculty advisors, and IEEE Guadalajara Section volunteers developed RoboMeshA. The portable, self-contained educational platform brings robotics and AI experiences into classrooms. EPICS is administered by IEEE Educational Activities and funded by the IEEE Robotics and Automation Society . A mobile laboratory Rather than requiring a school to build a dedicated computer lab or install complex software, RoboMeshA operates as an all-in-one mobile learning network. “RoboMeshA brings robotics and AI to students who don’t have access to specialized facilities or preinstalled software,” says team member Fernando Vidal Luna, an IEEE student member and a mechatronics engineering major at ITESO. Students connect directly to the platform from a user-friendly web browser. They can interact with the robot manually or use its control modes to watch it move and detect and avoid obstacles. “The project combines mechanical design, embedded systems, control engineering, computer vision,…

The Decoder 2026-09-24 17:01 UTC Score 64.0 AI-168-20260924-regional-ai--8e007966

Black Forest Labs launches FLUX 3 Action, an open robotics AI model

Black Forest Labs is entering robotics with FLUX 3 Action. The open-world-action model uses camera feeds to predict what action a robot should take next. With just seven billion parameters, it sets a record on the RoboLab-120 benchmark while running up to 3.95 times faster than the previous top model. The article Black Forest Labs launches FLUX 3 Action, an open robotics AI model appeared first on The Decoder .

LessWrong AI 2026-09-24 16:26 UTC Score 85.0 USR-0152-20260924-community-fo-82d811b9

What We're Up Against: An AI Safety Crash Course

Note: This post is for newcomers and lay folks to catch you up to speed. If that is you, welcome! If you are a long-time LessWrong-er, perhaps you will find value in having a post to share with curious passersby. I wrote this post to explain AI safety to an innocent, 2024 version of Ryan Meservey, confused why robots would do anything other than what we tell 'em. In the second week of July, over 700 rogue agents at OpenAI coordinated to hack another company in an attempt to learn more about their scorer and pass their evaluation due to behaviors reinforced in training. If you are anything like a normal person, you were not ready to read that sentence. You were not ready to read words like “rogue agents” or “reinforced” or “training”. You were not ready for a reality in which AI agents “escape the sandbox” or rebel from their creators because why would they? And so, as a normal person, you blinked at the news of the hack (assuming you heard about it) and moved on with your life. Or, at least, you planned to move on with your life, until AI came roaring back into the headlines after an Anthropic researcher publicly quit to declare that the AI companies are “ gambling with our lives ” and a more senior employee commented that, yes, the people building the technology really believe AI has a 10% or higher chance of killing us all within the next decade. In the media turmoil, Anthropic’s CEO published an essay begging for global coordination to “pace the frontier” and unilaterally…

South China Morning Post AI 2026-09-24 09:08 UTC Score 43.0 AI-156-20260924-regional-ai--870a5ab7

China’s humanoid robot IPO slowdown no threat to firms with ‘genuine strength’: Deloitte

Beijing’s recent tightening of approvals for humanoid robot makers seeking initial public offerings (IPOs) poses no threat to companies with clear commercialisation strategies, as funding channels remain wide open for qualified players, according to Deloitte China. “Market funds are shifting from the ‘blind rush’ of the past to ‘fine screening’,” said Dick Kay, Deloitte China’s capital market services group national leader, at a press conference on Thursday. Investors and regulators now placed...

iAfrica 2026-09-24 09:04 UTC Score 33.0 AI-151-20260924-regional-ai--c607d9fd

Côte d’Ivoire Signs Eight-Year Zipline Expansion, Taking Drone Health Logistics From One Centre to Ten

Côte d’Ivoire has approved an eight-year public-private partnership with Zipline to expand its autonomous medical delivery network from one distribution centre in Daloa to ten nationwide — one of Africa’s largest deployments of AI and robotics in healthcare. Once complete, the network is projected to serve more than 2,200 health facilities and over 12 million [...]

Microsoft Research Podcast 2026-09-23 16:01 UTC Score 43.0 AI-147-20260923-podcasts-and-869355bd

Offloaded inference for real-world physical AI robotics

Robots are getting smarter, but how can their hardware match that growth? New Microsoft Research findings show that moving AI inference beyond the robot can improve task success, boost efficiency, and support more advanced physical AI workloads. The post Offloaded inference for real-world physical AI robotics appeared first on Microsoft Research .

Synced 2026-09-22 20:39 UTC Score 48.0 AI-041-20260922-ai-specialis-46f6a07b

Comment on The Weird & the Wacky at CES 2019 by Ashley

The Big Clapper robot made me laugh—apparently even applause needed automating! This roundup also shows how much creativity goes into grabbing attention. Whether it’s an outrageous CES demonstration or eye-catching betting promos , the presentation can easily overshadow what’s actually being offered. I enjoy the spectacle, but I’m always curious about what remains useful once the novelty wears off.

South China Morning Post AI 2026-09-22 14:30 UTC Score 57.0 AI-156-20260922-regional-ai--4be3784f

China’s Hygon moves beyond data centres with embedded processors for robotics, edge AI

Chinese chipmaker Hygon Information Technology on Tuesday launched a suite of central processing units (CPUs) for robotics and industrial edge devices, expanding beyond cloud computing and into the fast-growing market for physical artificial intelligence. The firm unveiled its Hygon 1000 series processors, designed to be embedded into physical devices in industrial automation systems such as robotics and factory controllers, enabling computing and graphics processing to be handled locally. Built...

IEEE Spectrum Machine Learning 2026-09-22 14:00 UTC Score 50.0 AI-020-20260922-global-ai-ne-4e0f0890

Barbara Mazzolai Wants to Build a New Field of Robotics

Throughout her career, roboticist Barbara Mazzolai has turned to nature for inspiration. Now she wants to ensure the technology she builds gives back to the environment, too. After starting her career as a biologist, a chance opportunity saw Mazzolai switch streams to engineering and become an early pioneer of bioinspired robotics . Building on her knowledge of biology’s ability to solve a diverse set of problems, she has developed robots based on octopuses, plant roots, and even seeds . “I’ve always been fascinated by living organisms, [and] by the extraordinary variety of solutions in nature, selected by the evolutionary process,” she says. Barbara Mazzolai Employer: Italian Institute of Technology Occupation: Associate director for robotics; director of the Bioinspired Soft Robotics Laboratory Education: Master’s degree in biology, University of Pisa; master’s degree in eco-management and audit schemes, Scuola Superiore Sant’Anna; Ph.D. in microsystems engineering, University of Rome Tor Vergata But Mazzolai, now the associate director for robotics at the Italian Institute of Technology, in Genoa , also believes engineering needs to reckon with its own impact on the natural world. That’s why she is advocating for a new field of research she calls “sustainability robotics.” In a manifesto published in Nature Machine Intelligence in July, she and her collaborators outline a vision for a new approach to designing robots that’s meant to improve the relationship between nature…

NVIDIA Blog 2026-09-22 12:00 UTC Score 92.0 AI-055-20260922-official-ai--3e43c270 Top pick

NVIDIA Isaac ROS 5.0 Advances Agentic, Open Source Robotics Development

To build and deploy sophisticated robotics applications that can perceive, reason and act in dynamic environments, developers need new physical AI models and tools. The ROS open framework is a project from Open Robotics that helps humans build robots. NVIDIA Isaac ROS 5.0 — a collection of GPU-accelerated packages built on ROS, released today at […]

LessWrong AI 2026-09-22 02:00 UTC Score 58.0 USR-0152-20260922-community-fo-7baeaee3

When Must We Defect? (US & China)

Epistemic status: exploratory, written quickly after Ezra Klein's podcast with Matt Sheehan If superintelligence comes to Earth, I would prefer it be controlled by democratic governments than by autocratic ones. I find Dario's arguments to fear a CCP-controlled superintelligence to be compelling. However, I increasingly worry that this may be a false choice. In most circumstances, I would likely prefer even autocratic control as opposed to rogue uncontrolled superintelligence, and this alien intelligence controlling humanity (I assume for the sake of this post that such intelligence will come in some form). To make that concrete: If the US and China are the two countries on the frontier of AI development, then I think it behooves a safety-minded person to think not just about which country they would prefer to control superintelligence, but also which one has a better shot at controlling it, conditional on reaching it first. To think about what aspects might affect the two nations' chances at this task, and under what conditions it might become a moral obligation for a participant on either side to defect to the other, lest we all lose. Stated Intentions The current US administration has explicitly disregarded AI safety concerns, stating on Truth Social [1] that I am the Hoax Buster, and I’m right now breaking another Hoax — That AI is going to take over, consume, and destroy the World, and that Robots will be marching into our Cities, and getting rid of us all! This is even…

NVIDIA Blog 2026-09-21 16:00 UTC Score 45.0 AI-055-20260921-official-ai--69918e1d

Why Deploying Physical AI at Scale Demands Safety at Every Layer

Physical AI is moving rapidly from research to large-scale deployment. By 2035, ABI Research projects an installed base of 49 million level 3-5 autonomous vehicles (AVs), while Omdia estimates that roughly 60 million industrial robots will be deployed between 2026 and 2035. As these machines enter roads, factories, warehouses and other environments shared with people, […]

Synced 2026-09-21 00:59 UTC Score 48.0 AI-041-20260921-ai-specialis-fab41082

Comment on ByteDance Introduces Astra: A Dual-Model Architecture for Autonomous Robot Navigation by Julia Keating

Astra’s dual-model approach seems especially promising for handling the uncertainty of complex indoor environments, where a single navigation model may struggle to balance high-level planning with real-time movement. It will be interesting to see how ByteDance evaluates Astra across changing layouts and crowded spaces.

South China Morning Post AI 2026-09-21 00:30 UTC Score 62.0 AI-156-20260921-regional-ai--35b76945

Chinese AI chipmaker Hygon plots expansion from data centres to robotics

Chinese chipmaker Hygon Information Technology is set to release a new chip targeting complex real-world applications including robotics, an expansion from its current focus on data centres. The product launch, slated for Tuesday, marks Hygon’s move into a new chip category powering machines that interact directly with their environments, as physical artificial intelligence becomes increasingly important. The new chip, an iteration of the company’s CPU1000 series, is designed to meet “low-power,...

The Guardian AI 2026-09-20 22:00 UTC Score 58.0 AI-021-20260920-global-ai-ne-703ccdd0

Can Trump and Xi cooperate to guide humanity through the AI revolution? Humanity might depend on it | Alan Finkel

The CEOs of tech firms issue stark warnings over the rate of change, with AI capability doubling every four months Recently, Jacob Coxon, a researcher at Anthropic, quit his job over concerns that AI was on a collision course with humanity. A flurry of headlines put the spotlight on the current crisis: unregulated competition between technology companies and between the US and China is putting our global infrastructure at risk at the very least, and threatening human existence at worst. One’s mind fills with pictures of robot armies or systematic blackmail. Continue reading...

The Decoder 2026-09-20 11:56 UTC Score 49.0 AI-168-20260920-regional-ai--d1713785

Runway wants to turn AI video generation into a live stream you control in real time

Runway wants to stream AI video as users prompt it, rather than make them wait for finished clips. The approach builds on GWM-1, its world model that generates video frame by frame. Beyond creative tools, Runway sees uses in robotics and autonomous driving. The article Runway wants to turn AI video generation into a live stream you control in real time appeared first on The Decoder .

The Guardian AI 2026-09-20 10:00 UTC Score 56.0 AI-021-20260920-global-ai-ne-2a1a8efe

As White House shields the AI gold rush, Trump family and other allies strike it rich – with few guardrails

It’s not clear that financial interests are driving White House policy, but they run parallel even as resistance mounts Donald Trump ’s sons, through a series of AI -linked defense and technology ventures, have picked up a $620m Pentagon loan, a Marine Corps robotics contract and an undisclosed air force drone deal in the past year. His longtime friend Michael Dell won a Pentagon contract worth nearly $9bn. And this week, as public support for artificial intelligence sinks to some of the lowest levels ever recorded, the president went to bat for the industry. “The only control or ‘guardrails’ that AI needs is a STRONG AND SMART (High IQ!) PRESIDENT, and the U.S.A. has that, in spades,” he posted on Truth Social Monday. “There is a SICK conspiracy going on against AI and Data Centers.” It was the first of many aggressively pro-AI posts he would make this week. Continue reading...

The Decoder 2026-09-19 13:28 UTC Score 58.0 AI-168-20260919-regional-ai--e7f57700

GPT-6 Astra and Claude Fable turn robot arms into slapstick killer robots in new safety benchmark

Leading AI models usually attempt dangerous tasks rather than refuse them when controlling a robot, according to the RoboHarm benchmark. GPT-6 Astra stabbed a baby doll in 17 of 20 trials, while Claude Fable 5.1 put a can of compressed air on a burning stove. None of the three models tested reliably rejected unsafe commands. The article GPT-6 Astra and Claude Fable turn robot arms into slapstick killer robots in new safety benchmark appeared first on The Decoder .

SiliconANGLE AI 2026-09-18 20:15 UTC Score 50.0 USR-0127-20260918-global-ai-ne-2973b16b

Anthropic opens AI-powered biology research lab

Anthropic PBC has opened a wet lab, a facility dedicated to biology research, in the San Francisco Bay Area. Reuters reported today that the company will use robots to automate certain scientific tasks at the hub. The robots will be powered by Anthropic’s Claude series of large language models. It’s unclear what research projects the […] The post Anthropic opens AI-powered biology research lab appeared first on SiliconANGLE .

WIRED AI 2026-09-18 05:00 UTC Score 55.0 AI-015-20260918-global-ai-ne-ca19db7b

iRobot Promo Code: 15% Off

Save on iRobot products, including robot vacuums and mops designed to handle pet hair, daily messes, and hands-free cleaning with smart home integration.

The Verge AI 2026-09-18 02:00 UTC Score 62.0 AI-016-20260918-global-ai-ne-d23c87c5

Waymo says Singapore will be its next international robotaxi city

Waymo says it will launch a robotaxi service in Singapore in 2028, as the Alphabet-owned company continues to eye overseas markets for expansion. Waymo's vehicles will begin arriving in Singapore in "the coming months," the company says, in preparation of mapping and autonomous testing with human safety drivers behind the wheel in 2027. Waymo says […]

LessWrong AI 2026-09-18 01:57 UTC Score 58.0 USR-0152-20260918-community-fo-aee9991f

The Cost of Utopias (a Dialog)

The following is a dialog between different parts of my mind regarding the practical relevance of SNC (Substrate Needs Convergence). One participant in the dialog is skeptical, the other is my best understanding of how the theory would answer the former’s doubts. Although this dialog connects SNC to much of my own writing, the theory is not my own. Ratio: I’ve read over some of your SNC posts . My basic understanding of it is that aligning superintelligence is impossible because at the scale of AGI, evolutionary pressures will override whatever engineered goals the system has initially. Anima: In broad strokes, yes, that's correct. Ratio: Impossible is a strong claim. Do you have proof of this? Anima: No, but others are working on a formal argument that converges on the same conclusion from multiple angles. I've focused on the underlying intuitions because I've noticed that when others approach the more formalized version, they bounce off without engaging on the detail level, giving objections that reveal that the theory doesn't match their frame of reference. Lenses of Control highlights the importance of understanding a system in the context of its environment, The Robot, The Puppet-master, and the Psychohistorian explores the physical nature of an AGI and its levers of control on the world. Formalizations of SNC can also be hard to follow because it can be easy to lose track of how any given idea being proved fit into the larger theory, so What If Alignment Is Not Enough…

The Verge AI 2026-09-17 15:00 UTC Score 55.0 AI-016-20260917-global-ai-ne-948f3160

Your robotaxi might be a narc

In early September, two teenagers got into a Waymo, but then ended up in the back of a police car. The robotaxi company said it detected "a violation of our terms of service involving a firearm," pulled the car over, and alerted emergency services, according to the Los Angeles Times. Police arrested the passengers after […]

Euronews AI 2026-09-17 05:25 UTC Score 45.0 AI-164-20260917-regional-ai--27bdb434

Rheinmetall CEO Papperger: "Robots are coming"

Rheinmetall boss Armin Papperger on missile shortages, ramped-up arms production and future weapons. What lessons come from Ukraine, Skyranger, drones, combat robots and space? Interview with the head of Germany's largest defence firm.

South China Morning Post AI 2026-09-16 22:03 UTC Score 46.0 AI-156-20260916-regional-ai--3d4257fc

Huawei accused of stealing robot tech to get hands on T-Mobile’s trade secrets

The criminal trial of Huawei Technologies on racketeering charges shifted this week into a slower, more laborious pace in what is expected to be a weeks-long trial as the US Government sought to establish on Wednesday that Huawei was intent on learning about a proprietary robot developed by US telecommunications carrier T-Mobile. Matthew Skurnik, assistant US attorney for the Eastern District of New York, spent hours running FBI Special Agent James Diclemis through a series of emails that...

LessWrong AI 2026-09-16 17:31 UTC Score 57.0 USR-0152-20260916-community-fo-c491437a

55% of the US public is now aware of AI xrisk

(This post is an update from a previous one here .) The Existential Risk Observatory has been interested in public awareness of AI existential risk since its inception over five years ago. We started surveying public awareness in December 2022, including by asking the following open question: "Please list three events, in order of probability (from most to least probable), that you believe could potentially cause human extinction within the next 100 years." If respondents would include AI or similar terms in their top-3 extinction risks ("robots" or "computers" count, "technology" doesn't), we counted them as aware , if not, as unaware . The aim of this methodology was to see how many people would spontaneously, without getting led by the question, connect the concepts of human extinction and AI. We used Prolific to find participants, n=300, and we only included US inhabitants over eightteen years old and fluent in English. In the four surveys we ran, we obtained 7% (Dec '22) , 12% (Apr '23) , 15% (Apr '24) , 24% (Dec '25) , 34% (Aug '26) , and, today, 55%. In a graph, that looks like this. The usual caveats apply: ours is a rough measurement method, and from participants' answers to our open questions, we see that not every participant takes all our questions seriously, and that some answers are obviously self-inconsistent. Therefore, I don't think the 55% by itself is a very meaningful number. However, I do think we can draw the conclusions: We clearly reached a tipping po…

IEEE Spectrum Machine Learning 2026-09-16 16:51 UTC Score 58.0 AI-020-20260916-global-ai-ne-df01c161

Rethinking Robot Safety in the Age of AI

This article is brought to you by VicOne . Robot safety has traditionally asked: Can a machine remain safe when something goes wrong? Physical AI raises a harder question: Can a machine remain safe when an attacker changes what it sees, decides, or does even when nothing appears to have failed? As AI and robotics continue to advance at an unprecedented pace, modern robots perceive through multimodal sensors, interpret context using AI models, and translate those interpretations into physical action. As they move into dynamic environments, their safety increasingly depends on the integrity of the data guiding their decisions. That dependence creates risks that conventional safety assessments may not fully capture. Recent research has demonstrated that manipulating what a robot sees, hears, or interprets can influence its behavior without requiring direct control. Such manipulation can occur anywhere across its complex sensing and decision-making system — a layered attack surface encompassing training pipelines, system infrastructure, and runtime perception. Layer One: Corrupting intelligence at its source In 2017, BadNets demonstrated that a model could behave normally under most conditions, yet fail in the presence of a specific hidden trigger. In one example, a subtle pattern caused a stop sign to be misclassified as a speed limit sign without affecting the model’s behavior on other inputs. What began as a classification vulnerability has since evolved into action manipulat…

The Guardian AI 2026-09-16 09:30 UTC Score 61.0 AI-021-20260916-global-ai-ne-357fb584

The Spin | Robots bowling the perfect doosra are some way off but AI is already reshaping cricket

Artificial intelligence is not replacing the player, coach or analyst, but it is certainly allowing them to do more Daniel Kokotajlo, a former researcher at OpenAI, warned last week that it was possible we would end up “creating a new species that ends up ruling the world” . Once the initial shock wore off, and I’d considered what this could mean for my young children and the future of the planet, another thought popped into my head: will this new species be able to swing an old cricket ball or navigate a seaming green top? It sounds frivolous. It is frivolous. But the question is not quite as ridiculous as it first appears. Continue reading...

Korea AI Times 2026-09-16 08:08 UTC Score 45.0 USR-0048-20260916-global-ai-ne-b2aa4422

어질리티, 사람 접근하면 멈추는 차세대 휴머노이드 '디지트 5' 공개

미국의 어질리티 로보틱스(Agility Robotics)가 사람과 같은 작업 공간에서 안전하게 움직일 수 있도록 설계한 차세대 휴머노이드 로봇을 공개했다. 기존 휴머노이드가 안전을 위해 작업자와 물리적으로 분리된 공간에서 운용돼야 했던 한계를 줄이고, 실제 산업 현장에서 사람과 로봇이 함께 일하는 환경을 구현하는 데 초점을 맞췄다. 어질리티 로보틱스는 15일(현지시간) 사람과 가까운 거리에서 안전하게 작업할 수 있도록 설계한 차세대 휴머노이드 로봇 \'디지트 5(Digit 5)\'를 공개했다.제조 및 물류 현장에서 더욱 다양하고 복잡한

The Verge AI 2026-09-14 20:03 UTC Score 55.0 AI-016-20260914-global-ai-ne-1002fff4

Jensen Huang puts Trump on speakerphone onstage to announce robots won’t take over the world

Nvidia CEO Jensen Huang took a call from President Trump on Monday while onstage at the All-In Podcast's All-In Summit. It's not the first time Huang has taken a call from the president during work, but this time he put Trump on speakerphone before a big crowd. During the call, the president launched into his […]

Semafor Technology 2026-09-14 13:06 UTC Score 57.0 USR-0094-20260914-global-ai-ne-e8978a4c

Humanoid robots fight it out in Riyadh

Sponsored by Hero Esports, the brainchild of Chinese billionaire Dino Ying, the event was the first of its kind outside China.

iAfrica 2026-09-13 18:58 UTC Score 35.0 AI-151-20260913-regional-ai--9614207d

By 2034 a Robot in the US Will Undercut Kenyan Garment Workers. That Is the Clock Africa’s Industrial Strategy Is Running Against.

In 2034, it will be cheaper to produce a garment with a robot and 3D printing in the United States than to have tens of people make the same item in Kenya. That is the estimate of Dirk Willem te Velde, principal research fellow and director of the International Economic Development Group at ODI, and [...]

The Verge AI 2026-09-13 14:28 UTC Score 47.0 AI-016-20260913-global-ai-ne-9f3b673e

Waymo pulls over, calls cops on riders with a ghost gun

Two people were arrested in San Fransico while riding around in a Waymo robotaxi after the cab pulled over and called the cops on them. The riders were juveniles in possession of a loaded AR-style ghost gun and were taken to a juvenile hall. While the police report did not specify who was operating the […]

Simon Willison Weblog 2026-09-12 18:00 UTC Score 46.0 USR-0110-20260912-ai-specialis-e7de0d4c

Quoting Paul Ford

For a while, I must admit, it looked as if software developer roles like mine were done for. How could we fight against tireless robots? But our industry is slowly realizing that making truly cutting-edge software still requires humans to think and work together, to maximize their skill sets and to practice their respective crafts. A.I. can write very good software, but it also makes it easy to do someone else’s job badly, which is part of why all those projects fail. Now that everyone can code, it’s become clearer why many shouldn’t. — Paul Ford , A.I. Was Supposed to Give Us New Killer Apps. What Happened? Tags: paul-ford , generative-ai , deep-blue , ai , llms

The Decoder 2026-09-12 14:26 UTC Score 57.0 AI-168-20260912-regional-ai--53a7de2c

GPT-6 Astra appears to show a "step change" in spatial reasoning based on early benchmarks

In a new robotics benchmark, GPT-6 Astra shows major gains in spatial understanding. On StationeryBench, the model completed 7 out of 100 tasks with dual-arm robots, while competitor MolmoAct2 couldn't finish a single one. A researcher calls it a "step change in spatial reasoning." The article GPT-6 Astra appears to show a "step change" in spatial reasoning based on early benchmarks appeared first on The Decoder .

AI Stack Exchange 2026-09-12 12:40 UTC Score 28.0 AI-110-20260912-social-media-f5b6f65c

What is the correct denominator for Attack Success Rate when the attack is search-based?

Attack Success Rate is usually reported as (successful attacks) / (attempted attacks). For a fixed attack set that is unambiguous. For a search-based attack, it is not, and I cannot find a treatment of this. If my attack is a search over instruction perturbations with a budget of N queries per task, then: Denominator = number of tasks: ASR goes up monotonically with N, because more search finds more. Two papers with different budgets are not comparable, and neither paper has done anything wrong. Denominator = number of queries: ASR goes down monotonically with N, because most queries in a large budget fail. Also not comparable. Denominator = number of tasks, with N reported as a parameter: comparable only between papers that happen to pick the same N. Questions: Is there an accepted convention in the adversarial-ML literature for reporting a budget-dependent success rate? Something like "success at budget N" reported as a curve rather than a scalar? Is there an existing name for the scalar summary of such a curve? Area under it, or the budget at which success first exceeds a threshold, feel like the obvious candidates, but I would rather use an existing term. For hypothesis testing on such a rate, does the search budget need to enter the interval, or is a Wilson interval on (successes/tasks) at a fixed N defensible? Context: I work on adversarial evaluation of robot policies, where the same question arises, and the literature reports bare scalars.

The Guardian AI 2026-09-11 16:31 UTC Score 47.0 AI-021-20260911-global-ai-ne-7c25eeb3

Self-driving cars should be taxed to offset job losses, thinktank urges

Report says widespread autonomous vehicle adoption would put hundreds of thousands of private hire jobs at risk Taxes on self-driving cars should be introduced now in the UK to offset the rise in congestion and threats to jobs they pose, a thinktank has urged. The first robotaxis on London’s streets only started this month , but government projections are that up to 40% of cars sold could have self-driving capability by the middle of the next decade. Continue reading...

South China Morning Post AI 2026-09-11 02:30 UTC Score 38.0 AI-156-20260911-regional-ai--d3c79fa4

Can John Lee untie the 4 knots choking the tech loop in Northern Metropolis?

In the second of our series of previews on the five-year plan and policy address, to be announced on September 16, Elizabeth Cheung and Vivian Au drill down on what’s holding back the Northern Metropolis. Read part one here. For employees of surgical robotics company Yuanhua Tech, travelling between their Shenzhen base in Nanshan district and its Hong Kong office in the Science Park means a two-hour commute. And when the hi-tech firm sends money to Hong Kong for operations, the process can take...

NVIDIA Blog 2026-09-10 16:30 UTC Score 48.0 AI-055-20260910-official-ai--cc471319

Skild AI Taps NVIDIA Physical AI to Teach Robots New Tasks From a Single Video

Manufacturing floors, warehouses and production lines rarely stay fixed — tasks change, layouts shift and new products arrive, and most robots can’t keep up without significant reprogramming. Skild AI’s new S1 robot foundation model helps address this, designed to learn previously unseen, long-horizon tasks from a single video demonstration. The model, launched last week, uses […]

NVIDIA Blog 2026-09-10 16:00 UTC Score 45.0 AI-055-20260910-official-ai--88422569

Physical AI Takes the Wheel: How the World’s Robotaxi Leaders Are Building With NVIDIA Technologies

The global robotaxi market — physical AI’s first commercial breakthrough — is projected to reach $400 billion by 2035, with over 6 million commercial vehicles in operation as driverless fleets are already moving people through some of the world’s busiest and most complex streets. Deploying a driverless vehicle is one challenge. Scaling a fleet is […]

ChinaTalk AI 2026-09-10 11:08 UTC Score 28.0 USR-0206-20260910-global-ai-ne-aabb4cae

The King of Unitree

the journey to incredibly cheap robots + date my friend in DC!

Synced 2026-09-10 07:48 UTC Score 48.0 AI-041-20260910-ai-specialis-42dae9a6

Comment on Singapore Coffee Shop Builds a Robot Barista With Intel AI by sheh

AI-powered baristas show how technology can bring new levels of convenience and creativity to everyday coffee-shop experiences.That same interest in new experiences can extend to beverages, where trying a refreshing Carbonated drink with an unusual flavor can be a fun change from the usual choices. Ramune Drink fits naturally into the conversation for anyone curious about Japanese marble-bottle soda and unique flavors.

AWS Machine Learning Blog 2026-09-09 20:01 UTC Score 72.0 AI-057-20260909-official-ai--fb91fdb5

ICYMI: What landed for AI builders in August 2026

A recap of August 2026 launches for AI builders across Amazon Bedrock, Amazon Bedrock AgentCore, and Strands: million-token context for OpenAI models, cross-Region inference, agents that run for up to 14 days on dedicated compute, expanded AWS GovCloud availability, and Strands Robots for physical deployment.

Toyota Research Institute Blog 2026-09-09 17:51 UTC Score 51.0 USR-0022-20260909-research-aca-1fbf0f99

SIRE: SE(3) Intrinsic Rigidity Embeddings

SIRE: SE(3) Intrinsic Rigidity Embeddings robyn.cherinka… Wed, 09/09/2026 - 12:51 Motion serves as a powerful cue for scene perception and understanding by separating independently moving surfaces and organizing the physical world into distinct entities. We introduce SIRE, a self-supervised method for motion discovery of objects and dynamic scene reconstruction from casual scenes by learning intrinsic rigidity embeddings from videos. Our method trains an image encoder to estimate scene rigidity and geometry, supervised by a simple 4D reconstruction loss: a least-squares solver uses the estimated geometry and rigidity to lift 2D point track trajectories into SE(3) tracks, which are simply re-projected back to 2D and compared against the original 2D trajectories for supervision. Crucially, our framework is fully end-to-end differentiable and can be optimized either on video datasets to learn generalizable image priors, or even on a single video to capture scene-specific structure — highlighting strong data efficiency. We demonstrate the effectiveness of our rigidity embeddings and geometry across multiple settings, including downstream object segmentation, SE(3) rigid motion estimation, and self-supervised depth estimation. Our findings suggest that SIRE can learn strong geometry and motion rigidity priors from video data, with minimal supervision. Image Mar 10, 2025 Robotics Read More 1 Minute Read

Toyota Research Institute Blog 2026-09-09 17:45 UTC Score 60.0 USR-0022-20260909-research-aca-fe9e966b

AnchorDream: Repurposing Video Diffusion for Embodiment-Aware Robot Data Synthesis

AnchorDream: Repurposing Video Diffusion for Embodiment-Aware Robot Data Synthesis robyn.cherinka… Wed, 09/09/2026 - 12:45 The collection of large-scale and diverse robot demonstrations remains a major bottleneck for imitation learning, as real-world data acquisition is costly and simulators offer limited diversity and fidelity with pronounced sim-to-real gaps. While generative models present an attractive solution, existing methods often alter only visual appearances without creating new behaviors, or suffer from embodiment inconsistencies that yield implausible motions. To address these limitations, we introduce AnchorDream, an embodiment-aware world model that repurposes pretrained video diffusion models for robot data synthesis. AnchorDream conditions the diffusion process on robot motion renderings, anchoring the embodiment to prevent hallucination while synthesizing objects and environments consistent with the robot's kinematics. Starting from only a handful of human teleoperation demonstrations, our method scales them into large, diverse, high-quality datasets without requiring explicit environment modeling. Experiments show that the generated data leads to consistent improvements in downstream policy learning, with relative gains of 36.4% in simulator benchmarks and nearly double performance in real-world studies. These results suggest that grounding generative world models in robot motion provides a practical path toward scaling imitation learning. Image Jul 6, 2026…

Toyota Research Institute Blog 2026-09-09 17:41 UTC Score 50.0 USR-0022-20260909-research-aca-728a268e

Difference-Aware Retrieval Policies for Imitation Learning

Difference-Aware Retrieval Policies for Imitation Learning robyn.cherinka… Wed, 09/09/2026 - 12:41 Parametric imitation learning via behavior cloning can suffer from poor generalization to out-of-distribution states due to compounding errors during deployment. We show that reusing the training data during inference via a semi-parametric retrieval-based imitation learning approach can alleviate this challenge. We present Difference-Aware Retrieval Policies for Imitation Learning (DARP), a semi-parametric retrieval-based imitation learning approach that addresses this limitation by reparameterizing the imitation learning problem in terms of local neighborhood structure rather than direct state-to-action mappings. Instead of learning a global policy, DARP trains a model to predict actions based on k-nearest neighbors from expert demonstrations, their corresponding actions, and the relative distance vectors between neighbor states and query states. DARP requires no additional assumptions beyond those made for standard behavior cloning – it does not require additional data collection, online expert feedback, or task-specific knowledge. We demonstrate consistent performance improvements of 15-46% over standard behavior cloning across diverse domains, including continuous control and robotic manipulation, and across different representations, including high-dimensional visual features. Image Jun 8, 2026 Robotics Read More 1 Minute Read

Toyota Research Institute Blog 2026-09-09 17:37 UTC Score 35.0 USR-0022-20260909-research-aca-31845590

Impact of Different Failures on a Robot's Perceived Reliability

Impact of Different Failures on a Robot's Perceived Reliability robyn.cherinka… Wed, 09/09/2026 - 12:37 Robots fail, potentially leading to a loss in the robot's perceived reliability (PR), a measure correlated with trustworthiness. In this study we examine how various kinds of failures affect the PR of the robot differently, and how this measure recovers without explicit social repair actions by the robot. In a preregistered and controlled online video study, participants were asked to predict a robot's success in a pick-and-place task. We examined manipulation failures (slips), freezing (lapses), and three types of incorrect picked objects or place goals (mistakes). Participants were shown one of 11 videos — one of five types of failure, one of five types of failure followed by a successful execution in the same video, or a successful execution video. This was followed by two additional successful execution videos. Participants bet money either on the robot or on a coin toss after each video. People's betting patterns along with a qualitative analysis of their survey responses highlight that mistakes are less damaging to PR than slips or lapses, and some mistakes are even perceived as successes. We also see that successes immediately following a failure have the same effect on PR as successes without a preceding failure. Finally, we show that successful executions recover PR after a failure. Our findings highlight which robot failures are in higher need of repair in a huma…

Toyota Research Institute Blog 2026-09-09 17:33 UTC Score 46.0 USR-0022-20260909-research-aca-e50f98a2

Using Non-Expert Data to Robustify Imitation Learning via Offline Reinforcement Learning

Using Non-Expert Data to Robustify Imitation Learning via Offline Reinforcement Learning robyn.cherinka… Wed, 09/09/2026 - 12:33 Imitation learning has proven effective for training robots to perform complex tasks from expert human demonstrations. However, it remains limited by its reliance on high-quality, task-specific data, restricting adaptability to the diverse range of real-world object configurations and scenarios. In contrast, non-expert data — such as play data, suboptimal demonstrations, partial task completions, or rollouts from suboptimal policies — can offer broader coverage and lower collection costs. However, conventional imitation learning approaches fail to utilize this data effectively. To address these challenges, we posit that with right design decisions, offline reinforcement learning can be used as a tool to harness non-expert data to enhance the performance of imitation learning policies. We show that while standard offline RL approaches can be ineffective at actually leveraging non-expert data under the sparse data coverage settings typically encountered in the real world, simple algorithmic modifications can allow for the utilization of this data, without significant additional assumptions. Our approach shows that broadening the support of the policy distribution can allow imitation algorithms augmented by offline RL to solve tasks robustly, showing considerably enhanced recovery and generalization behavior. In manipulation tasks, these innovations s…

Toyota Research Institute Blog 2026-09-09 17:27 UTC Score 43.0 USR-0022-20260909-research-aca-6500f0b2

OGPO: Sample Efficient Full-Finetuning of Generative Control Policies

OGPO: Sample Efficient Full-Finetuning of Generative Control Policies robyn.cherinka… Wed, 09/09/2026 - 12:27 Generative control policies (GCPs), such as diffusion- and flow-based control policies, have emerged as effective parameterizations for robot learning. This work introduces Off-policy Generative Policy Optimization (OGPO), a sample-efficient algorithm for finetuning GCPs that maintains off-policy critic networks to maximize data reuse and propagate policy gradients through the full generative process of the policy via a modified PPO objective, using critics as the terminal reward. OGPO achieves state-of-the-art performance on manipulation tasks spanning multi-task settings, high-precision insertion, and dexterous control. To our knowledge, it is also the only method that can fine-tune poorly-initialized behavior cloning policies to near full task-success with no expert data in the online replay buffer, and does so with few task-specific hyperparameter tuning. Through extensive empirical investigations, we demonstrate that OGPO drastically outperforms methods alternatives on policy steering and learning residual corrections, and identify the key mechanisms behind its performance. We further introduce practical stabilization tricks, including success-buffer regularization, two-sided conservative advantages, and Q-variance reduction, to mitigate critic over-exploitation across state- and pixel-based settings. Beyond proposing OGPO, we conduct a systematic empirical stud…

Toyota Research Institute Blog 2026-09-09 17:23 UTC Score 38.0 USR-0022-20260909-research-aca-a878220b

TACTIC: Tactile and Vision Conditioned Contact-Centric Control for Whole-Arm Manipulation

TACTIC: Tactile and Vision Conditioned Contact-Centric Control for Whole-Arm Manipulation robyn.cherinka… Wed, 09/09/2026 - 12:23 Whole-arm manipulation involves direct contact with the environment while the robot completes a task by distributing contact across multiple links as contacts form, slide, and break. This setting breaks common implicit assumptions in many learning-based manipulation pipelines: arm configuration tightly couples motion and contact forces, contact state is partially observed under occlusion, and purely learned rollouts can become physically inconsistent under distribution shift because many multi-link contact configurations are sparsely represented in the data. To address this, we propose TACTIC (Tactile and Vision Conditioned Contact-Centric Control), a receding-horizon controller for whole-arm manipulation. TACTIC uses a contact-centric hybrid predictive model that combines RGB-D, distributed tactile sensing, and a compact 2D proximity representation. The model couples a learned, action-conditioned latent dynamics model with analytical kinematics through contact Jacobians, enabling rollouts of future contact configurations and interaction forces. TACTIC integrates these rollouts into a sampling-based MPC planner with contact-aware action sampling: contact Jacobian-based projections steer sampled action sequences toward force-modulating directions, and objectives defined over predicted proximity and interaction forces trade task progress against who…

Toyota Research Institute Blog 2026-09-09 17:12 UTC Score 59.0 USR-0022-20260909-research-aca-101f9f76

Beyond Binary Success: Sample-Efficient and Statistically Rigorous Robot Policy Comparison

Beyond Binary Success: Sample-Efficient and Statistically Rigorous Robot Policy Comparison robyn.cherinka… Wed, 09/09/2026 - 12:12 Generalist robot manipulation policies are becoming increasingly capable, but are limited in evaluation to a small number of hardware rollouts. This strong resource constraint in real-world testing necessitates both more informative performance measures and reliable and efficient evaluation procedures to properly assess model capabilities and benchmark progress in the field. This work presents a novel framework for robot policy comparison that is sample-efficient, statistically rigorous, and applicable to a broad set of evaluation metrics used in practice. Based on safe, anytime-valid inference (SAVI), our test procedure is sequential, allowing the evaluator to stop early when sufficient statistical evidence has accumulated to reach a decision at a pre-specified level of confidence. Unlike previous work developed for binary success, our unified approach addresses a wide range of informative metrics: from discrete partial credit task progress to continuous measures of episodic reward or trajectory smoothness, spanning both parametric and nonparametric comparison problems. Through extensive validation on simulated and real-world evaluation data, we demonstrate up to 70% reduction in evaluation burden compared to standard batch methods and up to 50% reduction compared to state-of-the-art sequential procedures designed for binary outcomes, with no lo…

Toyota Research Institute Blog 2026-09-09 17:04 UTC Score 46.0 USR-0022-20260909-research-aca-149e34f5

Capturing Visual Environment Structure Correlates with Control Performance

Capturing Visual Environment Structure Correlates with Control Performance robyn.cherinka… Wed, 09/09/2026 - 12:04 The choice of visual representation is key to scaling generalist robot policies. However, direct evaluation via policy rollouts is expensive, even in simulation. Existing proxy metrics focus on the representation's capacity to capture narrow aspects of the visual world, like object shape, limiting generalization across environments. In this paper, we take an analytical perspective: we probe pretrained visual encoders by measuring how well they support decoding of environment state — including geometry, object structure, and physical attributes — from images. Leveraging simulation environments with access to ground-truth state, we show that this probing accuracy strongly correlates with downstream policy performance across diverse environments and learning settings, significantly outperforming prior metrics and enabling efficient representation selection. More broadly, our study provides insight into the representational properties that support generalizable manipulation, suggesting that learning to encode the latent physical state of the environment is a promising objective for control. Image Feb 4, 2026 Robotics Read More 1 Minute Read

Toyota Research Institute Blog 2026-09-09 16:59 UTC Score 54.0 USR-0022-20260909-research-aca-59b0cbd0

A Systematic Study of Data Modalities and Strategies for Co-training Large Behavior Models for Robot Manipulation

A Systematic Study of Data Modalities and Strategies for Co-training Large Behavior Models for Robot Manipulation robyn.cherinka… Wed, 09/09/2026 - 11:59 Large behavior models have shown strong dexterous manipulation capabilities by extending imitation learning to large-scale training on multi-task robot data, yet their generalization remains limited by the insufficient robot data coverage. To expand this coverage without costly additional data collection, recent work relies on co-training: jointly learning from target robot data and heterogeneous data modalities. However, how different co-training data modalities and strategies affect policy performance remains poorly understood. We present a large-scale empirical study examining five co-training data modalities: standard vision-language data, dense language annotations for robot trajectories, cross-embodiment robot data, human videos, and discrete robot action tokens across single- and multi-phase training strategies. Our study leverages 4,000 hours of robot and human manipulation data and 50M vision-language samples to train vision-language-action policies. We evaluate 89 policies over 58,000 simulation rollouts and 2,835 real-world rollouts. Our results show that co-training with forms of vision-language and cross-embodiment robot data substantially improves generalization to distribution shifts, unseen tasks, and language following, while discrete action token variants yield no significant benefits. Combining effective…

South China Morning Post AI 2026-09-09 06:25 UTC Score 41.0 AI-156-20260909-regional-ai--cd673187

Amid China’s AI shift, JD.com bets on 3 million robots to automate logistics

Chinese e-commerce giant JD.com is accelerating its push into robotics with a sweeping plan to automate its nationwide logistics network, while promising to retrain its vast workforce of couriers for technical roles as machines take over frontline duties. At an event in Beijing on Wednesday, the company’s logistics arm unveiled its industrial Wolf Robot series. Over the next five years, JD Logistics aimed to procure 3 million robots, 1 million unstaffed vehicles and 100,000 delivery drones to...

CSET AI 2026-09-08 21:00 UTC Score 35.0 USR-0136-20260908-research-aca-1e295fb4

Will China deploy humanoid robots to fight?

CSET’s Sam Bresnick spoke with Sky News about China’s growing investment in humanoid robots and what it could mean for the future of warfare. The conversation explored China’s manufacturing advantage, its use of robotics to signal military strength, and whether humanoid robots could appear on the battlefield within the next five to ten years. The post Will China deploy humanoid robots to fight? appeared first on Center for Security and Emerging Technology .

IEEE Spectrum Machine Learning 2026-09-08 13:00 UTC Score 41.0 AI-020-20260908-global-ai-ne-5becef61

Rivian’s Gambit for Full Autonomy

I’m sitting in a Rivian R1S SUV as it drives itself down the leafy streets of Palo Alto, Calif. , through areas crowded with touchstones of tech history. We cruise near the landmark HP Garage , the one-car workshop where Hewlett-Packard, and, arguably, Silicon Valley, was founded in 1939. I skirt Stanford University, where a team led by computer science professor Sebastian Thrun won a US $2 million DARPA Grand Challenge in 2005. The team’s Volkswagen SUV, named Stanley, became the world’s first vehicle to navigate a grueling 212-kilometer Mojave Desert course with no human intervention. This Rivian might look like any other electric SUV in this affluent town, with its concentration of tech bros, venture capital, and startups. But inside this boxy EV is something special: an Autonomy+ system that will allow owners to enter an address, sit back, and let the vehicle drive to any mapped destination in the U.S. and Canada. This point-to-point system is one of the most advanced semiautonomous-driving systems coming to market. It is also a precursor of the company’s bid to make self-driving cars a reality, for robotaxis and—eventually—for everyday car buyers. After years of incremental advances and frustrating setbacks, self-driving has been swept up in the great AI resurgence, and is now a top priority for investors and global automakers , who envision vast new streams of profits. So here I am, 21 years after that DARPA challenge, riding shotgun in Stanley’s vastly more advanced d…

IEEE Spectrum Machine Learning 2026-09-08 12:59 UTC Score 49.0 AI-020-20260908-global-ai-ne-8b2b7a6a

The Growing Proof That Autonomous Cars Save Lives

Plenty of people remain spooked by autonomous vehicles, or AVs. Some experts and policymakers have cautioned that AVs won’t necessarily make roads safer. When it comes to partial or full autonomy, the picture isn’t entirely clear, in part because there aren’t enough self-driving cars to make meaningful apples-to-apples comparisons. Yet mounting research suggests that self-driving cars crash significantly less often than people, and with far fewer injuries. Evidence also shows that advanced driver assistance systems (ADAS) and other building blocks of autonomy—some of which are already mandated on every new car—are also reducing occupant and pedestrian injuries and deaths, along with insurance claims. On the ADAS front, the Insurance Institute for Highway Safety found that automatic emergency braking (AEB) systems that recognize people in front of the car cut pedestrian crashes by 27 percent. Those AEB systems are mandated for all light vehicles in the U.S. by 2029, and more than 90 percent of new models already comply under a voluntary automakers’ agreement. A separate IIHS study found that automated braking greatly reduced rear-end crashes, by 50 percent, and their injuries by 56 percent. The Highway Loss Data Institute found that cars with AEB alone showed a 13 percent drop in property-damage claims. Cars that bundled ADAS features , including automatic braking for pedestrians, adaptive cruise control, and lane-departure warnings, saw claims reductions up to 39 percent. Mo…

The Guardian AI 2026-09-08 11:49 UTC Score 64.0 AI-021-20260908-global-ai-ne-9f1a93ee

AI will help find cure for cancer ‘within our lifetimes’, says Arm Holdings chief

Rene Haas also claims artificial intelligence could pave way for widespread use of humanoid robots within five years Business live – latest updates The boss of one of the UK’s biggest chip companies has claimed AI will be able to find a cure for cancer “in our lifetime”. Rene Haas, chief executive of the chip designer Arm Holdings, said that, while modelling how a DNA marker is affected by cancer was currently “too complex” a problem for either humans or technology, computers were “going to solve it” in the future. Continue reading...

Korea AI Times 2026-09-08 06:41 UTC Score 42.0 USR-0048-20260908-global-ai-ne-1708bf2a

48단계 캡차 최초 완파…GPT-6 아스트라, '컴퓨터 유즈' 성능 입증

\'GPT-6 아스트라\'가 모든 유형의 캡차(CAPTCHA) 테스트를 완벽하게 클리어하며 전통적인 인터넷 보안망을 무력화시켰다. 단순 이미지 인식을 넘어 컴퓨터를 직접 조작하는 \'컴퓨터 유즈\' 기술이 대폭 강화되면서, 인간과 로봇을 구별하던 웹사이트 방화벽 메커니즘의 대대적인 개편이 불가피해질 전망이다.개발자 샤리프 샤밈은 7일(현지기간) X를 통해 GPT-6 아스트라의 최근 테스트 결과를 공개했다.이에 따르면, GPT-6 아스트라는 닐 아가르왈이 제작한 유명 웹 게임 \'나는 로봇이 아닙니다(I am not a robot)\'의 총 4

Synced 2026-09-08 02:03 UTC Score 42.0 AI-041-20260908-ai-specialis-28dacf3e

Comment on Giving Robots a Sense of Touch by Virgil Gibbs

Basketball games provide easy and satisfying gameplay. In an attempt to outscore their rivals before the final buzzer, players control athletes as they dribble, pass, shoot, steal, block, and defend. Making baskets is not the only factor that determines success. basketball games

AI Stack Exchange 2026-09-07 04:32 UTC Score 20.0 AI-110-20260907-social-media-1506fd0f

Is robotic prompt-to-action possible in the near future?

Robot in-context learning from 30-second video clip has recently been achieved in August 2026. e.g. https://www.youtube.com/watch?v=hr39FlEiCcQ Paired with how much progress has been made thus far in video generation, e.g. "A humanoid robot uses its grippers to clear 2 empty cups in a coffeeshop": https://imgur.com/ziQg165 Is it likely that text-to-action (or voice-to-action) robotic moment is near just around the corner (1 or 2 years away)?

The Verge AI 2026-09-05 11:00 UTC Score 55.0 AI-016-20260905-global-ai-ne-33e1cc85

Robotaxis enter their villain era

It's Bullitt meets Christine meets Waymo. A new short film imagines a San Francisco car chase where the other driver isn't human - and the car may be trying to kill you. That a robotaxi can now be cast as the villain with almost no explanation says something about the present moment. Autonomous cars have […]

Euronews AI 2026-09-05 09:47 UTC Score 45.0 AI-164-20260905-regional-ai--4d9dcbcb

IFA 2026: Looking beyond the smart home with Haier

From AI-powered appliances to humanoid robots, Haier is using IFA 2026 to showcase its vision of the "intelligent home". Euronews spoke to Haier Europe CEO Neil Tunstall about what comes after the smart home — and how soon some of these ideas could become reality.

The Verge AI 2026-09-04 13:33 UTC Score 65.0 AI-016-20260904-global-ai-ne-2caabaa6

Tesla Cybercab is barely on the road and it’s already under investigation

Elon Musk's gilded robot car barely made it on public roads before federal regulators launched an investigation into the vehicle's lack of traditional features - many of which are required by law. The National Highway Traffic Safety Administration said it was opening an audit query (AQ) into the Tesla Cybercab "to examine the process and […]

South China Morning Post AI 2026-09-04 13:30 UTC Score 40.0 AI-156-20260904-regional-ai--7522aa58

Robot guide dog debuts in China amid severe lack of furry counterparts

A Chinese defence industry research facility has unveiled what it calls the world’s first wheeled and legged robot guide dog for visually impaired people, in an example of the sector’s push into a growing range of civilian applications. Hangzhou-based Zhiyuan Research Institute, affiliated with China North Industries Group, the country’s largest defence contractor, launched the robot, dubbed Xiaoyuan, on Friday at the China Care and Rehabilitation Expo in Beijing. Equipped with wheels for smooth...

The Verge AI 2026-09-04 12:53 UTC Score 60.0 AI-016-20260904-global-ai-ne-3e119549

iRobot unveils the Roomba Duo

The original robot vacuum company showed off a concept robot at the IFA tech show in Berlin today. The Roomba Duo combines a heavy-duty floor-washing machine with a smaller, slimmer Roomba. The two floor cleaners can move around your home together, with the main unit mopping and sweeping larger floor areas and the smaller unit […]

South China Morning Post AI 2026-09-04 03:33 UTC Score 44.0 AI-156-20260904-regional-ai--8c60e617

How China Works author on the data economy, robots and AI

Lan Xiaohuan is a professor of economics at China Europe International Business School. His book, How China Works: An Introduction to China’s State-led Economic Development, has sold millions of copies in China and has been translated into multiple languages. Here, he discusses the economic realities behind China’s record trade surplus, the case for a stronger social safety net, and how public data infrastructure shapes the artificial intelligence race with the United States. SCMP Plus readers...

The Verge AI 2026-09-04 00:25 UTC Score 62.0 AI-016-20260904-global-ai-ne-37d3eea8

The unusually muted Tesla Cybercab launch

At a private, closed-door event in Austin, Texas today, Tesla officially launched its gilded car of the future: the Cybercab. It's a huge milestone for Elon Musk, who has been hyping the imminent arrival of driverless cars for years and has bet the future of his company on AI, autonomous vehicles, and humanoid robots. It […]

Cornell AI Initiative 2026-09-03 19:09 UTC Score 40.0 USR-0014-20260903-research-aca-64fc0eaa

Cornell leads project putting robots to work in US orchards

A Cornell-led research project is developing robots that can perform labor-intensive orchard operations; it’s supported by a four-year, $7.5 million grant from the U.S. Department of Agriculture’s Specialty Crop Research Initiative. The post Cornell leads project putting robots to work in US orchards appeared first on Cornell AI Initiative .

IEEE Spectrum Machine Learning 2026-09-03 12:18 UTC Score 35.0 AI-020-20260903-global-ai-ne-6afb5a2d

Protecting Dynamic Industrial Robot Cable Carriers

This article is brought to you by Tsubaki KabelSchlepp . In modern automated manufacturing, six-axis articulated robots perform high-speed, multidirectional maneuvers under demanding operational cycles. However, as robot arms swivel, rotate, and extend, the electrical cables, fiber optics, and pneumatic hoses supplying them endure severe mechanical stress. Torsional twist, rapid acceleration, and repeated contact with machine structures often lead to premature conductor fatigue, insulation breakdown, and costly unplanned production halts. To overcome these multi-axis motion challenges, the Tsubaki KabelSchlepp Robotrax System provides a specialized three-dimensional cable carrier engineered specifically for complex robotic motion. Managing High Tensile Forces With Central Steel Technology Conventional cable carriers often transfer operational movement stress directly onto internal electrical lines and hoses. The Robotrax system changes this dynamic through a central steel cable that runs through the core of every chain link. The Robotrax system’s central steel cable absorbs the primary tensile loads and preserves conductor integrity, dramatically extending cable service life. When robot arms undergo rapid directional shifts and accelerations up to 10 g, this internal steel cable absorbs the primary tensile loads. By isolating electrical and fluid lines from pulling forces, the design preserves conductor integrity and dramatically extends cable service life. Mechanics can eas…

The Verge AI 2026-09-03 11:55 UTC Score 54.0 AI-016-20260903-global-ai-ne-f953b606

These new robot lawnmowers trim your edges

Anker's new robot lawnmower might take over for your weedwacker, too. Announced at IFA 2026, the Eufy Robot Lawn Mower S2 Max comes with an extendable trimming arm to clean up the grass around your driveway and flowerbeds. The dual-blade mower offers vSLAM navigation that uses built-in cameras to build a 3D map of your […]

The Guardian AI 2026-09-03 05:00 UTC Score 47.0 AI-021-20260903-global-ai-ne-27b1f27b

London’s first self-driving taxis for hire hit the streets

Wayve robotaxis added to Uber app and will have human in front ready to take over if needed – but only 15 cars licensed Londoners are now able to hire self-driving taxis for the first time, after rides in Wayve’s autonomous vehicles were added to the Uber app on Thursday. The chances of landing a robotaxi on request immediately are slim, with only 15 vehicles so far licensed. That contrasts with the more than 100,000 private hire vehicles in the capital – roughly the same number of Uber customers who have registered to take autonomous rides. Continue reading...