What's new
The 30 most-recently-filed papers and most-recently-updated concept pages.
Papers
- tweet Disabling thinking and exposing deep_think as a tool leaks the internal CoT reasoning format
- arxiv AttenA+: Rectifying Action Inequality in Robotic Foundation Models
- arxiv Lucida: Parse, Generate, and Place for Composable Real-to-Sim Scene Modeling
- blog LightNav-0: Scaling Real2Sim2Real for Zero-Shot Generalist Navigation
- blog Atlas: A World Model for Spatial Intelligence
- arxiv Zetta ζ: An Efficient Closed-Loop Embodied Harness for Self-Evolving Physical Intelligence
- arxiv Act with Intent: Distilling Behavior Intent for Vision-Language-Action Models
- arxiv There and Back Again: Bidirectional Diffusion Bridges for Multimodality Translation
- arxiv Sliding-window beats linear attention
- blog Beyond Data Scaling: Representation-Centric Continued Pre-training for Vision-Language-Action Models
- github Microduck RL — RL training environments for a ~800 g open-source bipedal robot
- yt Microduck Sim 2 Real
- arxiv LaGSplat: Inferring Physics-Governed Interactive Simulation from Monocular Video Using Latent Lagrangian Gaussian Splatting
- arxiv TemporalFlow-VLA: Learning Physically Grounded Execution History for Long-Horizon Robot Manipulation
- arxiv PAWBench: How Far Are We from Probabilistically Aligned World Modeling?
- github Inspect Robots — An Open-Source Evaluation Framework for Physical AI (Robocurve)
- tweet Lerrel Pinto — in-context learning for robots is not that hard
- arxiv That was not what I was aiming at! Differentiating human intent and outcome in a physically dynamic throwing task
- arxiv LeVJEPA: Efficient & Scalable Video Pretraining without the Heuristics
- github OpenDM — DM0.5: An Open-World Foundation Model for General-Purpose Embodied Intelligence
- tweet LeRobot demos Claude Code operating SO-ARM101 zero-shot via Anthropic MHS — self-calibration to 4.1mm accuracy
- tweet Lumos MOS 2 — heavy-duty industrial humanoid AI worker
- tweet Cua open-sources Computer History — encrypted local cross-session memory for computer-use agents
- arxiv V-JEPA 2: Self-Supervised Video Models Enable Understanding, Prediction and Planning
- arxiv Behavior Prompting Policy: Demonstrations as Prompts for Manipulation
- arxiv Beyond Imitation: Self-Improving Robot Policies via Off-Policy Q-Planning
- arxiv WarpSAC: Towards the Pinnacle of Scalable Off-Policy RL by Rethinking Exploration and Exploitation
- arxiv StreamPI: Streaming Multimodal Temporal Modeling for Vision-Language-Action Models
- arxiv Zero-WAM: In-Context World-Action Modeling from Human Videos for Open-Ended Task Generalization
- blog CARGO: Physical AI for Industrial Package Stacking
Concept pages
- concept 4D Scene Generation
- concept Autoregressive Video Generation
- concept Camera-Controlled Video Diffusion
- concept Context Length / Quality Trade-off in Video Generation
- concept Open foundation-model releases
- concept Pose Estimation and Motion Capture
- concept Reasoning RL
- concept Representation Autoencoders
- concept Synthetic Training Data
- concept Thinking with Modalities
Want a digest? It’s not currently posted to Slack. If that becomes useful, we can add the weekly trend digest covered in the original plan.