Physical AI, robots, and autonomous systems
Inspired by Andrej Karpathy's AutoResearch, I built a system where Claude Code acts as an autonomous ML researcher on tabular binary clas...
The idea is called HALO - Hierarchical Autonomous Learning Organism. The core premise is simple: what if instead of just making LLMs bigg...
The idea is called HALO - Hierarchical Autonomous Learning Organism. The core premise is simple: what if instead of just making LLMs bigg...
![[2603.03295] Language Model Goal Selection Differs from Humans' in an Open-Ended Task](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
Abstract page for arXiv paper 2603.03295: Language Model Goal Selection Differs from Humans' in an Open-Ended Task
Abstract page for arXiv paper 2603.03800: A Rubric-Supervised Critic from Sparse Real-World Outcomes
Abstract page for arXiv paper 2603.03655: Mozi: Governed Autonomy for Drug Discovery LLM Agents
Abstract page for arXiv paper 2603.03456: Asymmetric Goal Drift in Coding Agents Under Value Conflict
A wrongful death suit against Google alleges that the tech giant's AI urged a Florida man to arm himself and steal a robot body for it.
I've been building a system to create physics-based humanoid characters in Unity that can learn through reinforcement learning -- and you...
Abstract page for arXiv paper 2601.12349: Zero-Permission Manipulation: Can We Trust Large Multimodal Model Powered GUI Agents?
Abstract page for arXiv paper 2601.20666: Learning Contextual Runtime Monitors for Safe AI-Based Autonomy
Abstract page for arXiv paper 2512.24075: Multi-Scenario Highway Lane-Change Intention Prediction: A Temporal Physics-Informed Multi-Moda...
Abstract page for arXiv paper 2506.06658: Self-Improving Loops for Visual Robotic Planning
Abstract page for arXiv paper 2511.09710: Echoing: Identity Failures when LLM Agents Talk to Each Other
Abstract page for arXiv paper 2510.05684: D2E: Scaling Vision-Action Pretraining on Desktop Data for Transfer to Embodied AI
Abstract page for arXiv paper 2412.00798: Combinatorial Rising Bandits
Abstract page for arXiv paper 2603.03278: Tether: Autonomous Functional Play with Correspondence-Driven Trajectory Warping
Abstract page for arXiv paper 2603.03280: How to Peel with a Knife: Aligning Fine-Grained Manipulation with Human Preference
Abstract page for arXiv paper 2603.02783: Generative adversarial imitation learning for robot swarms: Learning from human demonstrations ...
Abstract page for arXiv paper 2603.03074: Design Generative AI for Practitioners: Exploring Interaction Approaches Aligned with Creative ...
Abstract page for arXiv paper 2603.03024: MA-CoNav: A Master-Slave Multi-Agent Framework with Hierarchical Collaboration and Dual-Level R...
Abstract page for arXiv paper 2603.02623: Uni-Skill: Building Self-Evolving Skill Repository for Generalizable Robotic Manipulation
Abstract page for arXiv paper 2603.02553: Give me scissors: Collision-Free Dual-Arm Surgical Assistive Robot for Instrument Delivery
Get the latest news, tools, and insights delivered to your inbox.
Daily or weekly digest • Unsubscribe anytime