4 min read

🤖 Federal Robot Ban

Plus: Figure 03 Shows Off, Tokyo Enters Humanoid Race

Good Morning, Roboticists!

Generalist AI claims near-perfect task execution with minimal data. However?


US Plans Federal Robot Ban

TL;DR: U.S. lawmakers introduced the American Security Robotics Act, a bipartisan bill that would bar U.S. federal agencies from buying or using robots made by foreign adversaries like China. The proposal targets humanoid robots and unmanned ground systems, citing risks of hidden access points and remote control, while allowing limited exceptions for research and national security uses.


Figure 03 Shows Off Humanlike Skills

TL;DR: On the Shawn Ryan Show, Figure AI CEO Brett Adcock demonstrated the Figure 03 humanoid robot in a live walkthrough, where it walked smoothly, maintained balance, tracked objects, and shook hands with host Shawn Ryan. The AI-powered system performed humanlike movements and tasks in real time, leaving Ryan visibly surprised by its control and responsiveness.


Tokyo Robotics Enters Humanoid Race

TL;DR: Tokyo Robotics showcased a new bipedal humanoid prototype, demonstrating human-like walking, balance recovery under external pushes, and real-time full-body teleoperation controlled by a human operator. The system uses reinforcement learning trained in simulation, marking the company’s move into legged robotics as it continues developing toward more stable and autonomous operation.


Generalist AI Unveils GEN-1 Leap

TL;DR: Generalist AI launched GEN-1, a new model it claims reaches “robot mastery” with 99% task success rates and 3× faster performance, using just one hour of robot-specific data. Demonstrations showed long, reliable task execution and unexpected improvisation in real-world scenarios.


Agent Scaffolding Fixes Robot Control

TL;DR: Researchers from Nvidia, UC Berkeley, Stanford, and Carnegie Mellon introduced CaP-X, a framework testing how AI models control robots via self-written code. They found leading models fail without human-designed abstractions, but agentic scaffolding—like automated debugging, reusable functions, and parallel trial-and-error—can close the gap, with some setups reaching near human-level performance without additional training.


UK University Deploys Alan Robot

TL;DR: Durham University introduced “Alan,” a Unitree G1 Edu humanoid robot, as a shared research platform for AI, robotics, and human-robot interaction. Used by researchers and students, it will support studies on perception, decision-making, and real-world task execution, enabling experiments in everyday environments and advancing work in autonomous systems and assistive robotics.


LLMs Control Robots Via ROS

TL;DR: Researchers from Huawei Noah’s Ark Lab, TU Darmstadt, and ETH Zurich built an open-source system connecting large language models with the Robot Operating System, allowing robots to interpret natural-language commands and execute them as real-world actions. It converts instructions into step-by-step plans, supports code or behavior trees, and improves through feedback, enabling more flexible robot operation.


Qualcomm Backs Robotics Startups

TL;DR: Qualcomm has joined MassRobotics as a sponsor, bringing its Dragonwing robotics platform and newly launched Robotics Hub to support startups. The initiative offers tools, sample applications, and a collaborative space for developers to build and scale robotics systems. The partnership aims to accelerate real-world deployment of AI-driven robotics through shared infrastructure, community support, and industry collaboration.


READ MORE

Let the Future Come to Your Inbox

Stay ahead without drowning in information. We turn the most important signals across AI, tech, marketing, and future products into 5-minute reads you can actually finish.


TOGETHER WITH US

AI Secret Media Group is the world’s #1 AI & Tech Newsletter Group, reaching over 2 million leaders across the global innovation ecosystem, from OpenAI, Anthropic, Google, and Microsoft to top AI labs, VCs, and fast-growing startups.

We've helped promote over 500 Tech Brands. Will yours be the next?

Email our co-founder Mark directly at mark@aisecret.us if the button fails.