Researchers just introduced PaLM-E, an embodied multimodal language model that injects real-world sensor data directly into an LLM's embedding space. Scaling up to 562B parameters, it turns high-level prompts like "bring me the rice chips" into multi-step physical manipulation plans, even generalizing to objects the robot has never encountered before.
An AI agent just flew Lockheed Martin’s VISTA X-62A tactical aircraft for over 17 hours, marking the first time artificial intelligence has taken the stick on a tactical jet like this. The system mimicked a human pilot using new autonomous control algorithms, aiming to dramatically speed up how the military tests and matures uncrewed combat platforms.
UCLA researchers just developed a broadband diffractive optical processor that uses photons instead of electrons to run neural network computations. By processing thousands of light wavelengths simultaneously, it runs massive matrix operations at light speed with near-zero computing power consumption.
Scientists from the University of Stirling and Hack the Planet developed a smart camera trap that runs on-device AI and connects straight to satellites. Tested in Gabon, it identifies elephants with 82% accuracy and alerts rangers in seven minutes without Wi-Fi or cellular service, helping stop poachers and prevent human-wildlife conflict.
Roboticists like Columbia University's Hod Lipson are actively pursuing machine consciousness, once considered an academic taboo. Lipson defines consciousness as the ability to imagine oneself in the future to adapt and evolve, though critics warn the field is still plagued by fuzzy definitions and human anthropomorphism.
BMW just brought out Arnold Schwarzenegger at CES to unveil the i Vision Dee, an electric sports sedan designed around next-level human-machine interaction. While "Dee" stands for digital emotional experience, the standout feature is its electronic ink exterior, which can shift between patterns across 32 different colors.
Stanford researchers just developed a biocompatible spray-on smart skin that allows typing without a physical keyboard. Applied like sunscreen, the stretchable nanomesh senses electrical conductivity changes as fingers bend, while machine learning algorithms translate those subtle hand movements into letters on screen. Why it matters: It could pave the way for imperceptible wearable computing that washes right off.
Nvidia just rolled out a major update to its Isaac Sim robotics platform. The tool can now simulate human workers walking around warehouses and stacking boxes, giving robot AI a safe sandbox to practice avoiding collisions before hitting real factory floors. The release also brings RTX-simulated LiDAR alongside a new Isaac ORBIT tool for robotic arm manipulation.
Samsung just announced the Bespoke AI Wall Oven at CES 2023, packing an internal camera that can livestream cooking sessions directly to online audiences. The smart appliance recognizes up to 80 dishes, optimizes cooking conditions, and pings your phone before food burns. It launches in North America later in 2023, though pricing is still under wraps.
Princeton researchers just introduced ATLA, a method that uses GPT-3-generated descriptions of tool shapes and functions to train robotic arms. Providing language context allowed simulated arms to master unfamiliar tools significantly faster. Why it matters: large language models are proving to be an effective shortcut for giving physical robots common-sense adaptability.
