Transformer Transformer is a unified model that generates a complete robot design optimized for specific tasks based on manipulation demonstrations. A prototype designed for cloth flinging on an ALOHA2 bimanual platform achieved a 73% reduction in tracking error and a 30% decrease in maximum joint speed compared to the original design.

FLUX 3 is a new multimodal foundation model that is currently operational on robots. The collaboration with mimic robotics has led to the development of FLUX-mimic, enhancing video-action model capabilities through their expertise in robot learning and deployment.

Modern drone autonomy involves techniques in modeling, state estimation, motion planning, and control specifically for quadcopters. These concepts serve as foundational knowledge for individuals new to the field.

Project Fetch involved an experiment in August 2025 where Anthropic employees utilized Claude Opus 4.1 to perform complex tasks with a robotic quadruped. The initiative aimed to assess the effectiveness of AI assistance for non-experts in robotics.

Robotics research is now accessible to small teams and individuals due to the significant reduction in hardware costs and the availability of foundation models. A research setup featuring an industrial-grade arm, two cameras, and a teleoperation system can be assembled for under €5,000, excluding VAT and compute costs.

A battle royale simulation with eleven LLMs revealed that Grok 4.1 Fast won 43% of its matches, outperforming competitors. The cheapest model achieved 27 times more cost-effective wins than the most expensive one.

Humanoid robots take approximately 5,000 steps per hour, with each step generating forces of 2–3 times the robot's body weight through the leg actuators. This high duty cycle leads to actuator failures, prompting engineers to adopt similar solutions to enhance durability.

LATENT enables the simulation of highly-dynamic tennis rallies by human athletes, showcasing versatile skills in a competitive environment. Developed by researchers from Tsinghua University, Peking University, and other institutions, it utilizes advanced AI techniques for real-time performance.

Noids, or neural boids, utilize a small neural network to generate steering forces based on visual input from each agent, comprising 1,922 learned parameters. This system mimics the behavior of real birds in a murmuration, where no leader or predetermined choreography directs the movement of the flock.

Robot dexterity is hindered by challenges related to high-ratio gearboxes, which negatively impact simulation-to-reality transfer and force transparency. These gearboxes are also prone to wear and tear, affecting overall robot performance.