The Massachusetts Institute of Technology has come up what it thinks is a better way to teach robots general purpose skills. Derived from LLM techniques, the method provides robot intelligence access to an enormous amount of data at once, rather than exposing it to individual programs for specific tasks. Faster and more cost efficient, the approach has been referred to as a “brute force” approach to problem-solving, and machine learners have taken to it in lieu of individualized, task-specific “imitation learning.” Early tests show it outperforming traditional training by more than 20 percent under simulation and real-world conditions. Continue reading MIT Intros LLM-Inspired Teacher for General Purpose Robots

Robotics startup Figure AI — with investors including OpenAI, Nvidia and Microsoft — has released its next-gen humanoid, Figure 02. Its predecessor made a splash earlier this year with a demo that captured it conversing with an interlocutor as it organized household items and prepared a snack. Compared to the Figure 01 prototype, with exposed wiring and limited range of motion, Figure 02 is more polished. The latest iteration boasts skeletal improvements for heavier lifting as well as enhanced visual reasoning to assist with machine learning. The result is characterized as “a major leap” in AI-powered robotics, a category in which players include Tesla and 1X Technologies. Continue reading Humanoid Robot Figure 02 Touts Better Strength, Reasoning

Robotics firm Figure AI is getting a lot of attention for its humanoid robot, Figure 01, which the company unveiled along with news that it has raised $675 million, for a $2.6 billion valuation, from investors including OpenAI, Nvidia, Microsoft and Amazon founder Jeff Bezos. Pronounced “Figure One,” the general purpose robot looks and moves like a human, and can perform mundane tasks like serving food as well as undesirable jobs like picking up trash. It “sees” using “onboard cameras that feed into a large vision-language model (VLM) trained by OpenAI,” according to Figure co-founder and CEO Brett Adcock. Continue reading Figure Unveils Humanoid Robot, Draws Notable Investments

In the parade of attention-grabbing items queueing up for CES 2024, LG Labs’ “retro-futuristic concept speaker” is generating early buzz. The DukeBox combines old-timey vacuum tubes with a transparent OLED display that can be used to view programing or provide ambient imagery (of, say, a crackling fireplace that lets the tubes ghost through thanks to the OLED’s adjustable transparency). Also showcasing at the CES LG Labs zone: a biped robot equipped with AI, the Gram Fold 17-inch foldable OLED laptop and the CineBeam Qube 4K projector that displays up to 120-inch images. Continue reading LG DukeBox Speaker Touts Vacuum Tubes and OLED Display

Nvidia Research has debuted Eureka, an AI agent that autonomously teaches robots complex motor skills. Powered by OpenAI’s GPT-4, Eureka has successfully trained a robotic hand to handle a pen with the dexterity of a human — a first, according to Nvidia. Eureka has also enabled robots to do things like open drawers, manipulate scissors and toss and catch balls, along with dozens of other tasks. “Eureka is a first step toward developing new algorithms that integrate generative and reinforcement learning methods to solve hard tasks,” according to Nvidia Senior Director of AI Research Anima Anandkumar said. Continue reading Nvidia Leverages OpenAI’s GPT-4 to Train Dexterous Robots

Amazon now has more than 750,000 robots that relieve warehouse employees from some intensely repetitive tasks. Just in time for the holiday shopping rush, the company is adding a new robotic technology, Sequoia, which is already operating at a fulfillment center in Houston, Texas. Amazon says Sequoia is a complete “reimagining” of how inventory is stored and managed, resulting in a 75 percent improvement in how inventory is identified and housed, and also improves employee safety. Amazon is also expanding its robot workforce with the addition of a bipedal robot called Digit, from Agility Robotics. Continue reading Amazon Introduces New Robots to Improve Fulfillment Speed

“Robots are great specialists, but poor generalists,” according to Google DeepMind, which says models are typically trained for individual tasks, and changing a single variable can mean starting again from scratch. Now the London-based Alphabet subsidiary thinks it’s come up with a way to combine knowledge across robotics for a general purpose machine helper. In conjunction with 33 academic labs, Google DeepMind has pooled data from 22 different robot types to create the Open X-Embodiment dataset. Simultaneously, the group releases the RT-1-X robotics transformer (RT) model derived from RT-1. Continue reading DeepMind and Academics Advance General Purpose Robots

Meta Platforms has published a new AI technology, the Segment Anything Model (SAM) that the company claims can identify objects it hasn’t seen before. Acting on a text prompt, SAM will highlight items in a photo or video, picking out all the cats, for instance, or flowers. It can also execute other functions, such as generating a 3D construct using a single 2D image or extrapolating from things viewed in a mixed reality headset. Segment Anything can work in concert with other models, potentially minimizing the need for voluminous data sets for training. Continue reading New Meta AI Can Detect Objects It Has Not Been Trained On

Meta Platforms has shuttered its Connectivity division, which focused on areas including subsea cabling. The group’s functions will be divided among two other company units, Infrastructure and Central Products. Launched in 2013 as Facebook Connectivity, the business group developed connectivity technologies, concentrating on innovations like solar-powered drones, low-Earth orbit satellites and fiber-laying robots to expand the footprint of the company’s social platforms and other services. In 2021, it is estimated that more than 300 million people were able to access faster Internet services as a result of Meta Connectivity initiatives. Continue reading Meta Closes Its Connectivity Unit, Reassigns Responsibilities

Online game company and mobile app developer NetDragon Websoft has invested $40 million in Rokid, maker of 5G-friendly AR glasses for business applications. Both companies are based in China. NetDragon has been in the news this past month when it became the first company to appoint an AI as its “rotating CEO.” Following the Rokid announcement, it appears the firm may be interested in developing lifelike AI characters to inhabit its games and augment teaching and enhance its AR initiatives, though to hear NetDragon actual CEO, Liu Dejian tell it, the company can learn a lot from its new c-suite addition, Tang Yu. Continue reading Chinese Game Company Appoints AI CEO and Invests in AR

LIFE Productions producer Sam Khoze looked to cast a robot for his film at several robotics companies until he found Erica, the creation of Osaka University roboticist Hiroshi Ishiguro, who modeled her after images of Miss Universe finalists. Erica now stars in “b,” a $70 million feature film in pre-production. Erica possesses realistic human details, but one of her greatest strengths is immunity to COVID-19. Her walk and voice give her away as an android, so she will, therefore, perform most of her scenes sitting down. Continue reading Realistic Human Robot Is Featured in Upcoming Sci-Fi Movie

Amazon is working on a new wearable, codenamed Dylan, that reportedly can discern human emotions. The voice-activated gadget, developed by Amazon in collaboration with Lab126 and the Alexa voice software team, is worn on the wrist and is meant to address health and wellness. Lab126 previously worked with Amazon to build its Fire phone and Echo speaker. According to sources, the wearable includes microphones that pair with software and work with a smartphone app to glean the user’s emotional state via the sound of his/her voice. Continue reading Amazon Testing Wearable That Recognizes Your Emotions

According to sources, Amazon is developing a TiVo-like device, dubbed “Frank,” that records live TV. The sources added that the DVR device, which includes physical storage, connects to the company’s Fire TV box using the same wireless technology as the Echo. The Frank DVR also has the ability to record live TV and stream the video to a smartphone for later viewing, but Amazon, said the sources, isn’t certain it will debut the DVR with that feature, which is offered by TiVo and Dish Network’s Slingbox. Continue reading Amazon Is Reportedly Developing Live TV DVR to Rival TiVo

At Google’s I/O conference, chief executive Sundar Pichai reflected on the backlash against Silicon Valley companies while, at the same time, promoting the company’s advances and ambitions in artificial intelligence. Among those were specific positive solutions, such as an AI-powered software that helps diagnose eye disease, and a demonstration of what Google Assistant — in a variety of voices and accents — can do for ordinary consumers, and how Smart Compose in Gmail will suggest complete sentences to make the process speedier. Continue reading Google’s Third-Gen Tensor Processor Unit Key to AI Ambitions

Disney recently submitted a patent application for “a robot that will move and physically interact like an animated character.” The prototype sketch with the application shows a round body similar to the soft-robot character in Disney’s 2014 “Big Hero 6.” Before soft-body robots can be employed at Disney theme parks, the company will have to deal with the issue of robotic interaction safety, since the patent application implies that theme park guests will be able to hug the soft-body characters. Continue reading Disney Tests Prototype, Pursues Patent for Soft-Bodied Robot