What's new
The 30 most-recently-filed papers and most-recently-updated concept pages.
Papers
- arxiv Flow Matching in Feature Space for Stochastic World Modeling
- arxiv Scaling Behavior Foundation Model for Humanoid Robots
- arxiv Understanding Reasoning from Pretraining to Post-Training
- blog Happy Oyster — Real-Time World Model for Interactive Creation
- blog Introducing Cosmos 3 Edge
- tweet Pareto (Hebbian Robotics) indexes AllenAI's MolmoAct2 Bimanual YAM dataset — clusters + prompt normalization (Brandon Ong)
- tweet Danfei Xu — two root-level paradigm shifts in robot learning: Sim2Real for locomotion, Behavior Cloning for manipulation
- arxiv Learning to Fold: prizewinning solution at LeHome Challenge 2026 (1st place online, 2nd offline)
- arxiv Read It Back: Pretrained MLLMs Are Zero-Shot Reward Models for Text-to-Image Generation
- github RynnBrain 1.1 — Open Embodied Foundation Models (Alibaba DAMO Academy)
- tweet Qwen3.8 preview announcement — 2.4T-parameter frontier model, open-weights soon (Alibaba Qwen)
- arxiv FlashSAC: Fast and Stable Off-Policy Reinforcement Learning for High-Dimensional Robot Control
- arxiv Towards Predictive, Aligned, and Scalable Robot Learning (Lumo-2)
- tweet HandUMI — Robot-free bimanual data collection retargetable to any parallel-gripper arm (Brik H. Meza launch tweet)
- tweet Haiyu Wu — critique of reachability-augmented CEM planning in JEPA world models (with VISReg counter-example)
- tweet Sunday Robotics data engine — from one Memory Developer off Craigslist to 1,000+ (Perry Jia thread)
- yt EmoMime: Augmenting Social Behavior and Self-Expression via Wearable Robotic Limbs
- arxiv WAM-TTT: Steering World-Action Models by Watching Human Play at Test Time
- arxiv See like a Robot: Robot-Centric Pointmaps for Vision-Language-Action Models
- arxiv UniVR: Thinking in Visual Space for Unified Visual Reasoning
- arxiv DenseReward: Dense Reward Learning via Failure Synthesis for Robotic Manipulation
- arxiv Worlds in One Demo: A Synthetic Data Engine for Learning Open-World Mobile Manipulation
- blog HDR: Hierarchical Denoising for Multi-Step Visual Reasoning
- blog Mixture of Frames Policy: Multi-Frame Action Denoising for Bimanual Mobile Manipulation
- blog REGRIND: A Minimalist Retargeting-Guided RL Recipe for Dexterous Manipulation
- tweet Sunday Robotics finds on-device VLA inference beats cloud during ACT-2 eval (Cheng Chi)
- tweet Sunday Robotics ran the Chinchilla scaling law with their VLA model (Tony Zhao follow-up)
- arxiv E2HiL: Entropy-Guided Sample Selection for Efficient Real-World Human-in-the-Loop Reinforcement Learning
- arxiv RoboMemArena: A Comprehensive and Challenging Robotic Memory Benchmark
- arxiv UniIntervene: Agentic Intervention for Efficient Real-World Reinforcement Learning
Concept pages
- concept Autoregressive Video Generation
- concept Classifier-Free Guidance variants
- concept Diffusion Distillation
- concept Diffusion training efficiency
- concept Dual-stream diffusion transformer
- concept Human-to-Robot Retargeting
- concept Inference-Time Scaling
- concept Hyperparameter scaling laws
- concept Joint audio-video generation
- concept Open foundation-model releases
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.