0:43Same VLA, three data sources: Human Archive OTS wins on task success@babugi28 · 168 views · 2026-10-15VLAPre-training Data
1:34R2-WAM: What Visual Representations Do World Action Models Need?@ai_bites · 151 views · 2026-10-15World-action modelsVisual Representation
0:10Gemini 4 Argon Ties Opus 5.5 for SOTA on Real-Robot Manipulation@chooi_jeq · 263 views · 2026-10-14RobocurveBenchmarks
1:07SmolWAM: A World-Action Model Shrunk to 57ms on the Edge@robotsdigest · 219 views · 2026-10-13World ModelsEdge computing
1:43Gemini 4 Argon Builds Its Own Training Sims to Teach a Humanoid Cornhole@GoogleAI · 190 views · 2026-10-13HumanoidGemini
3:12Natura-1: A 50B MoE Foundation Model for General Dexterous Manipulation@naturewill · 282 views · 2026-10-13Foundation ModelsDexterous Manipulation
2:38Long-WAM: Scaling World-Action Model Context to 19.2s in Real Time@AaronWeiHuang · 498 views · 2026-10-08World-action modelsVLA
0:21SEAR: Benchmark for Self-Evolving LLM Agents on Robots@BangzhengL · 1.1k views · 2026-10-07LLMRobot Control
0:21Gemini Robotics-ER 2 with 3x FR3 tower build: guardrail rejects the colliding pose, agent keeps proposing@Official_m9g · 538 views · 2026-10-04Gemini RoboticsFranka