Home / AI

AI 119 articles

AI

ActSafeGuard Puts Hard Joint Limits Inside Flow-Matching VLAs

SJTU’s ray-scaling layer hits 100% step safety on π0.5 and Fast-WAM, then keeps a real ALOHA arm on a magnetic-ball track.

Robb Harlan
AI

FARM Reads Failure Out of a Frozen World Model

Robb Harlan
AI

FreqFM Conditions VLA Flow Matching on Action Frequency

Shar Hendrix
AI

ReactHuman Freezes the Scene and Asks the Model to Catch or Dodge

Robb Harlan
AI

2AM Keeps Task Memory in the Agent, Not the Robot Policy

Robb Harlan
AI

HuRo Paints a Robot Over 630,000 Human Videos

Robb Harlan
AI

IMLE-VLA Drops the 10-Step Action Head and Hits 55 Hz

Shar Hendrix
AI

MaP-WAM Remembers as a Plan, Not a Growing Video

Shar Hendrix
AI

UniMPA Grounds Predicted Futures in Actions the Robot Has Already Done

Robb Harlan
AI

GTA-2 Builds Robot Skills From Axes, Not End-to-End Actions

Shar Hendrix
AI

HaWMPO Down-Weights Hallucinated World-Model Rollouts

Robb Harlan
AI

JEPA Policy Skips Diffusion and Still Beats It on Five Robot Tasks

Robb Harlan
AI

Show-Harness Lets a VLM Play the Robot Like a Game

Shar Hendrix
AI

PPS Steers a Frozen π0.5 Without Touching Its Weights

Robb Harlan
AI

FailureSpot Marks the Instant a VLA Starts to Blow It

Shar Hendrix
AI

RoboSPA: Pick the Third-Farthest Bottle, Then Remember What You Hid

Shar Hendrix
AI

AGC-VLN Lets a Drone Draw the Map a Ground Robot Actually Drives

Robb Harlan
AI

GIFT Trains Robot Policies to Keep Geometry, Affordances, and Goals

Robb Harlan
AI

GPT-6 Astra Drops a Block in a Bowl 19 Times in 20. The Puzzle Still Stalls

Robb Harlan
AI

RoboTok Indexes Web Video by 3D Hand Trajectories, Not How the Kitchen Looks

Shar Hendrix
AI

AGIBOT Opens 11,430 Real Trajectories That Include the Failures

Shar Hendrix
AI

Astribot’s SmoothRL Fine-Tunes a VLA While the Robot Is Still Moving

Robb Harlan
AI

EGR Stops a VLA From Listening to the Wrong Camera

Robb Harlan
AI

FWBC-VLA Wipes a Board and Opens a Loaded Door Without a Force Sensor

Robb Harlan
AI

WISE Only Imagines When the Robot Is About to Touch Something

Shar Hendrix
AI

XR-2 Trains on 1,500 Hours of Two-Handed Housework, Then DAgger

Shar Hendrix
AI

Facet-0 Puts Force in the VLA — 82% on Sub-Millimeter PC Assembly

Robb Harlan
AI

Isaac Teleop Can Now Buzz a MANUS Glove Finger by Finger

Shar Hendrix
AI

Perceptron Opens Isaac 0.5, a 36B Robot Model

Shar Hendrix
AI

Skild S1 Learns a 10-Minute Job From One Video

Robb Harlan
AI

Anthropic’s Model Hardware Standard Lets Claude Drive Lab Robots

Robb Harlan
AI

CLAP Trains Video World Models Across Robot Bodies, Including G1

Robb Harlan
AI

Jetson Orin Nano 2: 78 TOPS, Same Board, Ships 2027

Robb Harlan
AI

Figure Opens Index: 16 Million Human Videos for Helix

Robb Harlan
AI

GVLA Stops Treating Every Gripper Like a Parallel Jaw

Shar Hendrix
AI

LD4WAM Learns Robot Actions From 5,000 Hours of Human Video

Robb Harlan
AI

ViTacPhys Estimates Mass and Stiffness Before It Squeezes

Robb Harlan
AI

GEN-1.5 Learns a New Robot Task From a 3-to-12-Second Demo

Shar Hendrix
AI

NVIDIA’s SONIC Turns Motion Tracking Into a Humanoid Foundation Model

Robb Harlan
AI

Illinois Fine-Tunes π0.5 With the Robot’s Own Failed Rollouts

Robb Harlan
AI

RoboEdit Turns Human Hand Videos Into 14 Million Robot Frames

Shar Hendrix
AI

Hydra-0 Treats Robot Actions as Pixel Motion

Robb Harlan
AI

τ0-VLA Thinks About the Next Subtask Before It Commits

Robb Harlan
AI

CoRe Recovers a Frozen VLA by Imagining the Next Move

Robb Harlan
AI

HAF Splits a Generalist VLA So a Humanoid Can Walk and Work

Robb Harlan
AI

ViTaR Lets a Frozen VLA Feel Contact Without Rewriting It

Shar Hendrix
AI

ART Lets a VLA Call Night-Vision Instead of Retraining

Robb Harlan
AI

BICPO-VLA Smooths the Handoff When the Next Chunk Is Late

Shar Hendrix
AI

ReflexVLA Chases Moving Cubes Instead of Frozen Scenes

Shar Hendrix
AI

DaViNCi Makes Street Navigation Continuous — and the Cars Move

Robb Harlan
AI

DreamX-Phi Predicts What a Bimanual Arm Will See Next

Shar Hendrix
AI

MiDAS Adapts a Generalist Robot Policy From One Demo

Shar Hendrix
AI

H2R-Bench: Pretty Robot Videos Still Fail the Transfer

Shar Hendrix
AI

Seeker Learns Where to Look From the Action Stream

Shar Hendrix
AI

Temporal GRPO Stops Punishing the Good Part of a Failed Rollout

Robb Harlan
AI

A Linear Probe Reads Task Progress Out of π0.5

Robb Harlan
AI

Galaxea G0.5 Puts Reasoning and Action in One Token Stream

Robb Harlan
AI

StellaVLA Adapts From One Demo by Teaching the Why

Shar Hendrix
AI

BooST Moves Franka Skills Onto a UR3 From Five Demos

Shar Hendrix
AI

FACT Trains World-Action Models on Failures Without Copying Bad Moves

Robb Harlan
AI

Flex-π Predicts RGB, Depth, and DINO Futures From One Checkpoint

Shar Hendrix
AI

Gated VLA-Cache Recomputes Only When the Policy Sounds Unsure

Robb Harlan
AI

Surgical WAM Turns Unlabeled Endoscopy Video Into Closed-Loop Control

Robb Harlan
AI

HarnessWAM Adds Planning Memory on Top of World Action Models

Shar Hendrix
AI

JEPA-WAM Couples Latent Transitions With Action Generation

Shar Hendrix
AI

RynnValue Turns Timestamps Into Robot Rewards at 7,000 Hours Scale

Robb Harlan
AI

SLIM-0.5B Packs Action-Grounded Latents Into a Compact Robot Policy

Robb Harlan
AI

Vid2WAM Distills Video Diffusion Into a Compact World-Action Student

Shar Hendrix
AI

XPolicyLab Wants One Adapter Per Robot Policy, Not One Per Benchmark

Shar Hendrix
AI

AtlasVLA Remembers the Room With Only a Wrist Camera

Shar Hendrix
AI

Cross-View Consistency Makes Flow VLAs Survive Camera Moves

Shar Hendrix
AI

CrossTracer: One Semantic Path, Residual Fixes per Robot

Shar Hendrix
AI

Dyna-2 Trains a World-Action Model on 1M Hours of Human Video

Robb Harlan
AI

LifelongCrossNav Keeps a 3D Map Across Floors and Goals

Shar Hendrix
AI

TEMPO Splits VLA RL: Freeze Semantics, Train Actions Fast

Robb Harlan
AI

Adaptive-WAM: Stop Generating the Future Video Just to Steer

Robb Harlan
AI

EvoHIL: Human-in-the-Loop RL That Survives a Lighting Change

Shar Hendrix
AI

GORDON Learns Object-Centric Rewards From Action-Free Video

Robb Harlan
AI

PanoVLA Puts 360° Context Into Mobile Bimanual Manipulation

Shar Hendrix
AI

SpaceVLA Puts User Grasp and Place Anchors Into OpenVLA

Shar Hendrix
AI

Zero-Shot VLM Controllers Often Aren't Looking at the Camera

Robb Harlan
AI

GUARD Watches Whether a Diffusion VLA Is Actually Looking at the Scene

Robb Harlan
AI

HiRoC Post-Trains VLAs as Planner Plus Subgoal Executor

Robb Harlan
AI

LiLa-WAM Trains a World-Action Model on One 24 GB GPU

Robb Harlan
AI

Mimir Splits World Memory from Task Memory for Long-Horizon Agents

Robb Harlan
AI

Robust-WAM Keeps Video Pretraining and Adds Semantic Foresight

Robb Harlan
AI

SpikingNav Navigates With Fewer FLOPs and Better Corruption Survival

Shar Hendrix
AI

DyPES-VLA: Shared Dynamics Priors, Native Actions Per Embodiment

Robb Harlan
AI

GeniWorld Turns Robot Actions Into Pictures the World Model Can Follow

Robb Harlan
AI

In-Context VLA: Consume Grounded Language, Don’t Narrate CoT

Robb Harlan
AI

JoyAI-RA 0.5: Scale Manipulation by Aligning Human Video to Robot Actions

Shar Hendrix
AI

SkillMemo Stores Atomic Skills and Pulls Them Back at Runtime

Robb Harlan
AI

VLAff Learns Where, How, and Which Way From Ego Video

Shar Hendrix
AI

W²-VLA Predicts How the Wrist Will Move Before It Moves

Shar Hendrix
AI

BridgeVLA++ Adds Spatio-Temporal Memory Without Blowing the Data Budget

Shar Hendrix
AI

SiMDex: Mine 5% of Ego Videos, Beat Random Mixing on Dexterous Tasks

Robb Harlan
AI

SARF Defends VLA Robots Against Physical Attention-Hijacking Patches

Shar Hendrix
AI

Track4Action Distills 3D World Tracking Into Tracker-Free VLA Policies

Shar Hendrix
AI

HERO Bootstraps Robot Skills From Zero Human Demos

Shar Hendrix
AI

HoloMotion: Horizon’s Open Whole-Body Humanoid Controller Hits 300 FPS

Robb Harlan
AI

HumanCLAW Asks: Can VLMs Act Through a Body?

Robb Harlan
AI

Robbyant Open-Sources LingBot-VLA 2.0 Across 20 Robot Morphologies

Shar Hendrix
AI

CLIFT Fine-Tunes Closed Gemini Robotics On-Device into Near-Perfect Humanoid Skills

Robb Harlan
AI

X Square Robot Opens Its Embodied AI Stack: Data, World Model, Action

Robb Harlan
AI

BrainCo Demos Thought-Controlled Robots at WAIC 2026

Shar Hendrix
AI

Humanoid’s KinetIQ Ascend Pushes Real-World RL Past Demo Manipulation

Robb Harlan
AI

LimX COSA 0.5: Whole-Body VLA That Tidies a Room Uncut

Robb Harlan
AI

RoboParty UFO: Open Unsupervised RL for Humanoid Control

Shar Hendrix
AI

Encord Is Tagging Brain Waves While Pilots Train Robot Arms

Shar Hendrix
AI

Flexion, Niantic Spatial, and NVIDIA Close a Sim2Real Loop for Humanoids

Robb Harlan
AI

Naver’s ARCBRAIN Flow Aims at No-Code Robot Service Design

Shar Hendrix
AI

Gemini Robotics 2: Whole-Body Control From Feet to Fingertips

Robb Harlan
AI

General Intuition Raises $320M to Train Agents on Gameplay Action Data

Shar Hendrix
AI

NVIDIA GR00T 1.7 and Isaac Teleop Land in Hugging Face LeRobot

Shar Hendrix
AI

Sanctuary AI Pivots Physical AI Onto Today’s Factory Arms

Shar Hendrix
AI

Generalist’s GEN-1 Learns Across Thousands of Robot Hands

Shar Hendrix
AI

Interlatent Opens a Low-Latency VR Teleop Stack (QUIC + Browser IK)

Shar Hendrix
AI

Enigma Raises $71M to Make Robot Control Feel Obvious

Shar Hendrix
AI

Sunday’s ACT-2: Memo Hits 99.1% Laundry Folds in Unseen Homes

Shar Hendrix