🤖 Robotics - High-Performance Marathon Robot
This article details Honor's robot Lightning, which surpassed the human half-marathon world record. It highlights the technical innovation responsible for its advanced performance.
Key Points:
• Robot Lightning completed a half-marathon in record time.
• Liquid cooling technology was essential to its performance.
• The system circulated over 4 liters of coolant per minute to its motors.
🔗 Resources:
• RoboDaily ↗ - Coverage of Honor's robot performance
Image
✨ AI Models - New Release Announcement
This article announces the upcoming public launch of new AI models, GPT-5.6 Sol, Terra, and Luna. Preview access is currently being expanded globally.
Key Points:
• GPT-5.6 Sol, Terra, and Luna models will launch publicly soon.
• Global preview access for these models is expanding.
🔗 Resources:
• OpenAI ↗ - Official announcement of new AI models
Image
🤖 Robotics - Contact-Rich Manipulation Challenges
This article discusses the ongoing challenges in contact-rich manipulation for robotics. It highlights the difficulty of tasks requiring repeated contact with robot hands using only vision.
Key Points:
• Contact-rich manipulation remains a significant challenge in robotics.
• Tasks like opening jars require complex multi-contact interactions.
• Pure vision alone is insufficient for these manipulation problems.
🔗 Resources:
• RoboPapers ↗ - Discussion on contact-rich manipulation in robotics
🚀 Agentic Robotics - Graph-as-Policy (GaP)
This article introduces Graph-as-Policy (GaP), a new approach in agentic robotics developed by NVIDIA and UC Berkeley. GaP uses computation graphs to manage complexity and improve interpretability.
Key Points:
• GaP is a new agentic robotics variant for managing complexity.
• It uses computation graphs to ensure modularity.
• GaP facilitates improved interpretability in robotic systems.
🔗 Resources:
• GaP Project Page ↗ - Open code and paper for Graph-as-Policy
Image
🤖 Robotics - Task and Motion Planning Survey
This article presents a new survey exploring the relationship between Task and Motion Planning (TAMP) and robot learning. It organizes current literature and introduces a taxonomy for TAMP learning.
Key Points:
• A new survey clarifies the connection between TAMP and robot learning.
• The survey organizes existing literature on TAMP and learning.
• It introduces a taxonomy for various TAMP learning approaches.
🔗 Resources:
• Yixuan Huang ↗ - Details on TAMP and robot learning survey
⭐️ Support
If you liked reading this report, please star ⭐️ this repository and follow me on Github ↗, 𝕏 (previously known as Twitter) ↗ to help others discover these resources and regular updates.