interestingengineering.com

AIAI-generated summary OpenAI plans to discontinue the Advanced and Standard voice options in ChatGPT on December 11, 2026, making GPT-Live the platform default. OpenAI says nearly all voice users have switched to GPT-Live since July, though live video and screen-sharing will temporarily keep using Advanced voice.

AIAI-generated summary Ant Group embodied AI company Robbyant has open-sourced LingBot-Touch, a tactile foundation model designed to give robots advanced touch perception using a dataset of nearly 5,000 hours of real-world tactile data. Robbyant says the model supports magnetic tactile, visual tactile, and piezoresistive sensing, which could improve fine manipulation tasks such as turning pages, unscrewing bottle caps, and identifying Mahjong tiles.

AIAI-generated summary XNRGY Climate Systems announced a single-month production output to help data centers handle intense thermal challenges from artificial intelligence infrastructure. The company shipped hardware delivering over 350 megawatts of cooling capacity during October from factories located in Mesa and Montreal.

AIAI-generated summary Google DeepMind reports that Gemini 4 Argon robotics research capabilities acted as a simulation engine to accelerate the development of Gemini Robotics 2.5. Google DeepMind says Argon generates simulated practice runs from camera scenes and builds digital twins to help robots generalize to new objects.

AIAI-generated summary Shield AI reports that its Hivemind autonomy software enabled two U.S. Air Force Collaborative Combat Aircraft to fly in formation and perform coordinated beyond-visual-range air combat maneuvers. The demonstration took place on October 26 at the Nevada Test and Training Range, where the uncrewed aircraft used the software to execute coordinated maneuvers against a virtual airborne threat.

AIAI-generated summary OpenAI introduced Control Room on November 9, 2026, as a central authorization layer for research workloads designed to cut off model output after safety or security interventions. OpenAI reports a maximum 35-second lag between receiving a stop instruction and further progress becoming impossible across its research infrastructure.

AIAI-generated summary Meta has launched its budget-tier artificial intelligence model named Muse Ember to compete with offerings from Anthropic and OpenAI. Cloudflare CTO Dane Knecht stated that Muse Ember ran up to 30 percent faster at the tail than the comparable model tested.

AIAI-generated summary Anthropic says it has upgraded Claude Fable and Claude Mythos to version 5.2, bringing improvements in coding, reasoning, and vision. Anthropic says the updated models generate outputs up to 30 percent faster than previous versions and are available at a lower price.

AIAI-generated summary Boston Dynamics released a video demonstrating its Atlas humanoid using wheels to move around during controls engineering experiments. Boston Dynamics stated that the wheeled Atlas is a research testbed rather than a product.

AIAI-generated summary At the Conference on Robot Learning in Austin, Texas, NVIDIA announced new open robotics technologies and models. NVIDIA says the offerings include GR00T 2 and Isaac Lab 3.0 to help researchers train robots for complex tasks.

AIAI-generated summary Alphabet Inc.'s Google is releasing new versions of its custom tensor processing units to compete with Nvidia Corp. in the artificial intelligence market. The latest chips are designed to support confidential computing and help models access a wider range of information for handling user queries.

AIAI-generated summary Swiss startup Embodied AI has introduced AIR, a two-armed mobile robot designed to operate alongside factory workers without safety cages. Embodied AI says the robot initially uses remote supervision before learning tasks autonomously within a month, achieving a 95 percent task success rate in paid pilots.

AIAI-generated summary Anthropic researchers used Claude to map a bacterial vault particle derived from a bacteriophage. Anthropic reports that its FlashProve system verifies model outputs using zero-knowledge cryptography without revealing proprietary model weights.

AIAI-generated summary AGIBOT deployed six G2 wheeled humanoid robots for a 24-hour convenience store shift in Shanghai to test embodied AI capabilities. AGIBOT says four machines passed their performance assessments after completing 1,730 tasks, while two robots failed the evaluation.

AIAI-generated summary Physical Superintelligence announced on November 6, 2026 that an internal research system running instances of a non-public Anthropic model found a proof of the Aaronson-Ambainis conjecture. Physical Superintelligence says the system produced a complete proof during a run on its infrastructure, and the Lean proof assistant accepted it.

AIAI-generated summary Xinhua and UBTECH Robotics have unveiled bionic humanoid news anchors at the World Media Summit in Beijing. Xinhua said that the robots will perform assisting roles to support editorial staff in the newsroom.

AIAI-generated summary MIT researchers have developed a memory system called UQ-DAAAM to help robots identify and correct inaccurate descriptions of objects stored in their memory. The technology evaluates multiple viewpoints and combines observations to reduce uncertainty, resolving ambiguities for approximately 92 percent of objects during testing.

AIAI-generated summary Nvidia Corp. is increasing the monitoring of artificial intelligence accelerators sold in Southeast Asia to combat semiconductor smuggling. The chipmaker now requires regional customers to use its fleet intelligence system to track hardware locations in order to become official Nvidia Cloud Partners.

AIAI-generated summary NVIDIA researchers report that their new visual planning method, VFlash, accelerates robot trajectory decoding by up to 12 times. The researchers state that VFlash reduces 3D trajectory prediction errors by as much as 61% in specific tests.

AIAI-generated summary Scientists at the Fraunhofer Institute for Digital Media Technology and Pan Acoustics developed a system that uses structure-borne sound sensors and machine learning to measure rainfall. The technology achieved an average error of 1.88 millimeters per hour during testing for rainfall rates between 2 and 5 millimeters per hour.