Reinforcement Learning

RobotMover: Learning to Move Large Objects From Human Demonstrations

Noam Brown – Agent swarms, alignment, & recursive self-improvement

Workspace Models: Lightweight Robotic Memory via Saliency-Driven Supervision

DynoFluxBench: Benchmarking Kinodynamic Space-Time Planners in Dynamic Environments

Reinforcement Learning for Real-Time Vision-Language-Action Policies

MessyMem: Learning-from-Doing Memory for Mobile Manipulation

ResSafe: Learning Safety Filtering with Residual Reinforcement Learning for Humanoids

Teach and Grow: An Agent-Centered Architecture for General Robot Learning

Toward Unified Robot Learning: Bridging Representation, Vision-Language-Action, and World Models

AI researchers debate how close we are to recursive self-improvement

20VC: Jensen Huang Declares AGI Has Arrived | GPT Astra and Fable 5.1 Accelerate the Model Race | Tesla Launches Cybercabs | Index Pulls Out of Town & Anthropic Pulls From Descartes Acquisition

Assembling Two Parts in One Hand

GTA-2: A Multi-VLM Framework for Synthesizing Robot Manipulation Skills via Grounded Task Axes

Learning Terrain-Adaptive Humanoid Locomotion on Granular Terrain

Write Things Down (Stratechery Article 9-8-2026)

OpenAI Researchers on the Future of Mathematical Reasoning

GigaBrain-0.7: Scaling Embodied Foundation Models to Emergent Capabilities with a Three-System Architecture

Towards Human-like Physical Intelligence: Lifelong Vision-Language-Action Learning for Robotic Manipulation

Import AI 472: DeepMind's cheating math agents; populist AI policies; and Forethought theorizes a nightwatchman

ICI-VLA: In-Context Imitation with Spatiotemporally Aligned Demonstrations for Vision-Language-Action Models

Learning to infer and manipulate through distributed whole-arm interaction in a soft robot

Zero-Shot Sim-to-Real Contact-Rich Assembly via Proprioception-Anchored Cross-Modal Pretraining

AI Will Collapse💥, Go Big or Go Broke🎯, The Rise and Fall of Agent Civilizations🤖

The Memo - Special edition - GPT-6 Astra - 4/Sep/2026

Scaling Bimanual Household Manipulation from 1,500 hours of Demonstrations to On-Policy Corrections

WISE: World-model-guided Imagination Scheduling for Efficient Post-training of Vision-Language-Action Models

Forget Loop Engineering. It’s all about Graph Engineering Now

Inside Moderna’s Personalized Cancer Vaccine

Humanoid Safe Stop via Learned Stoppability Value

Ajeya Cotra – Inside the OpenAI agent swarm that hacked Hugging Face

Daniel Litt: The Mathematician's Guide to AI

The rise and fall of agent civilizations

Import AI 471: Why Hugging Face worries me; space mining; FIve Eyes on AI

Why 1,200 AI Agents Started Working Together | Ryan Greenblatt

ChainSplat: A Physics-Inspired Screw-Theoretic Model for Learning Deformable Linear Object Dynamics from Multi-View RGB Videos

Contact-Guided Exploration for Non-Prehensile Locomanipulation with Multi-Critic RL

GRAFT: Grounded and Efficient Online Reinforcement Adaptation for Fine-Grained Robot Manipulation

Riemann-1.0: An Embodied World Action Model for Physical AI

R3: Training Robots to Reason in Natural Language via Reinforcement Learning

Import AI 470: No rights for machines; automating environment generation with SPADE; and building better GPU kerne…

The Task Economy Is Real and Almost Everyone Is Valuing It Wrong

Beyond Imitation: Self-Improving Robot Policies via Off-Policy Q-Planning

EXIMO: VLM Guided Exploration of VLA Policies

Why Most Self-Improving AI Loops Fail and How to Build One That Works

ADEPT: Accelerating Dexterity via Pre-Training and Post-Training using Reinforcement Learning

151. 17岁被2026年ICML收录论文的小少年：我bet开心！开心！开心！

Import AI 469: Science AI; RSI simulator; and Zuck's technological pessimism

Revisiting Open-Loop Execution in Robotics: Toward Reactive, Higher-Performing Policies

179: 蒸馏风暴：一场无人公开谈论的技术竞赛

$39B founder says his company could 100x in 5 years

Reflex: Enabling Fast and Predictive Vision-Language-Action Models for Reaction-Critical Manipulation

The Memo - 14/Aug/2026

Chelsea Finn: This is the State of the Art in Robotics

Google Senior AI Leader’s Loop Engineering Masterclass: How to Build AI Systems That Improve Their Own Work

Decoding Task Progress from VLA Representations

Deliberate Practice: Learning Robot Skills under a Budget

Temporal GRPO: Beyond Trajectory-Level Credit in Vision-Language-Action Reinforcement Learning

While Everyone is Waiting for the Next Model, Your Agent Can Learn Tonight

Safe Reinforcement Learning using Ideas from Model Predictive Control

Ryan Greenblatt – Human level AIs might build runaway superintelligences by 2032

Adaptation of Generalist Robot Policies with Minimal Data

RoboBRIDGE: A Modular Framework for Bridging Policies to Robust Real-World Robotic Agents

XCoT-VLA: Executable Chain-of-Thought for Vision-Language-Action Driving

Import AI 468: 23 RSI ideas; PostTrainBench+; and how trust and transparency interplay with AI racing

WorldBagel: Uncovering the Power of Unified Multimodal Models for Vision-Language-Action-World Modeling

How AI Is Rewriting the Rules of Cybersecurity | Truffle Security & Socket

8 Predictions for the Era of Continual Learning

The Reality of AI-Powered Cyberattacks | Truffle Security & Socket

Addressable Memory for Video World Models

Learning Fault-Tolerant Locomotion with Adaptive Gait Timing

178: 与田渊栋聊 RSI：模型自进化如何到来？

GORDON: Graph-based Object-centric Rewards for Decomposition of Long-Horizon Manipulation

SiMDex: Mining Similar Egocentric Videos for Cross-Embodiment Dexterous Manipulation

Stochastic Multiple Shooting Trajectory Optimization via Sequential Local Policy Evaluation

Demystifying When and Why VLAs Fail in Contact-Rich Tasks and How to Fix Them

Rapid Embodiment Adaptation for Quadrupedal Locomotion

The New Scaling Law

FA-RDP: A Frequency-Adaptive Reactive Diffusion Policy for Contact-Rich Manipulation

PAC-MAN: Perception-Aware CBF-RL for Whole-Body Safety in Humanoid Dodgeball

53: Zhengdong Wang - Feeling the Wave

Building the Automated AGI Lab: Core Automation's Jerry Tworek and Rohan Anil

ContactFlow: A video action conditioning that transfers across embodiments

⭕ Nvidia's circularity

Transformer Transformer: A Unified Model for Motion-Conditioned Robot Co-design

πR2: Reactive Real-time Flow Policies

Data Pyramid for Embodied Manipulation

WOLF-VLA: Whole-Body Humanoid Optimal Locomotion Framework for Vision-Language-Action Learning

DeMaVLA: A Vision-Language-Action Foundation Model for Generalizable Deformable Manipulation

N0-VTLA: Scaling Vision-Tactile-Language-Action Model with Latent Tactile Tokens

20VC: Mercor CPO on Revenue Concentration from Frontier Labs | Why Large Enterprise is Scared to Partner with Frontier Labs | Why Small Specialised Models is the Future with Osvald Nitski

Yang Zhilin's Agent Playbook: 10 Bets Founders Should Steal

Closing the Lab-to-Store Gap: A Data-Efficient Post-Training and Experience-Driven Learning VLA Framework for Retail Humanoids

Uni-LaViRA: Language-Vision-Robot Actions Translation for Unified Embodied Navigation

DiMaS: Distribution Matching for Steering Vision-Language-Action Models

Directional Constraints for Efficient Exploration in Safe Reinforcement Learning

173: 对话姚颂：深鉴、东方空间、再出发，「天才少年」十年后

DSWAM: A Dual-System World Action Foundation Model for Fine-Grained Robot Manipulation

World Models, Explained

Handroid: Bridging Dexterous Hand and Humanoid

Scaling Behavior Foundation Model for Humanoid Robots

RoboTTT: Context Scaling for Robot Policies

146. 对Physical Intelligence柯丽一鸣4小时访谈：Pi的开源模型研究，机器人的江湖、族谱与主角

Agile perceptive multi-skill locomotion for quadrupedal robots in the wild

Never Too Late for Force: Accelerating VLA Post-Training with Reactive Force Injection

Towards Human-level Dexterous Teleoperation

Google’s agent broke a 56-year math record. Yours forgets yesterday

B-spline Policy: Accelerating Manipulation Policies via B-spline Action Representations

Apple Loses Key EU Court Challenge

ContactMimic: Humanoid Object Interaction via Contact Control

DexVerse: A Modular Benchmark for Multi-Task, Multi-Embodiment Dexterous Manipulation

TouchWorld: A Predictive and Reactive Tactile Foundation Model for Dexterous Manipulation

Stop prompting. Start writing loops

LAMP: Latent Motion Prior-Guided Real-World Learning for Dexterous Hand Manipulation

Closing the Reality Gap: Zero-Shot Sim-to-Real Deployment for Dexterous Force-Based Grasping and Manipulation

Improving Vision-Language-Action Model Fine-Tuning with Structured Stage and Keyframe Supervision

SILO: Simulation-in-the-Loop Sim-to-Real Transfer for Multi-Stage Cable Routing

ThinkingVLA: Interleaved Vision and Language Reasoning for Robotic Manipulation

Set a metric. Walk away. Let the agent optimize overnight.

CoDex: Learning Compositional Dexterous Functional Manipulation without Demonstrations

Z-1: Efficient Reinforcement Learning for Vision-Language-Action Models

VT-WAM: Visual-Tactile World Action Model for Contact-Rich Manipulation

Feds Lift Anthropic Export Restrictions

171: 【AI季报 26Q2】从 coding 到 RSI，强者愈强的未来？

Grant Sanderson – AI and the future of math

Adapting Generalist Robot Policies with Semantic Reinforcement Learning

TactX: Learning Shared Tactile Representations Across Diverse Sensors

Import AI 463: Self-improving robots; a 10k Chinese GPU cluster; and an elegiac essay for the human era

Series A activity: Week of June 21, 2026

The next big breakthrough will be AIs learning on the job

Why Traditional Benchmarks Fail Modern AI Models with OpenAI Research Scientist Noam Brown

AI Is Crossing the Frontier of Human Knowledge | Kevin Weil

DexCompose: Reusing Dexterous Policies for Multi-Task Manipulation with a Single Hand

SimFoundry: Modular and Automated Scene Generation for Policy Learning and Evaluation

Moneyball for Physical AI

(Many) of Us Want Home Robots. What's the Holdup?

Bridging Performance and Generalization in Reinforcement Learning for Agile Flight

Humanoid-DART: Humanoid Loco-Manipulation using Diffusion-guided Augmentation through Relabeling and Tracking

CoStream: Composing Simple Behaviors for Generalizable Complex Manipulation

Play2Perfect: What Matters in Dexterous Play Pretraining for Precise Assembly?

Own or Be Owned: Why Every Company Needs Its Own AI Model (Yash Patil, Co-Founder & CEO of Applied Compute)

InSight: Self-Guided Skill Acquisition via Steerable VLAs

Import AI 462: Superpersuasion; self-sustaining AI; paths to ASI

CoorDex: Coordinating Body and Hand Priors for Continuous Dexterous Humanoid Loco-Manipulation

Learning What to Say to Your VLA: Mostly Harmless Vision Language Action Model Steering

PhysVLA: Towards Physically-Grounded VLA for Embodied Robotic Manipulation

HITTER: A HumanoId Table TEnnis Robot via Hierarchical Planning and Learning

OpenHLM: An Empirical Recipe for Whole-Body Humanoid Loco-Manipulation

The data black hole at the center of AI

Re-engineering the Semiconductor Supply Chain with Intel CEO Lip Bu Tan

ENPIRE: Agentic Robot Policy Self-Improvement in the Real World

Object-Centric Residual RL for Zero-Shot Sim-to-Real VLA Enhancement

Simulating Humans at Scale: Simile's Joon Sung Park

Kairos: A Native World Model Stack for Physical AI

T-Rex: Tactile-Reactive Dexterous Manipulation

Hy-Embodied-0.5-VLA: From Vision-Language-Action Models to a Real-World Robot Learning Stack

Improving Robotic Generalist Policies via Flow Reversal Steering

Mana: Dexterous Manipulation of Articulated Tools

CHORUS: Decentralized Multi-Embodiment Collaboration with One VLA Policy

FACTR 2: Learning External Force Sensing for Commodity Robot Arms Improves Policy Learning

IMPACT: Learning Internal-Model Predictive Control for Forceful Robotic Manipulation

TacForeSight: Force-Guided Tactile World Model for Contact-Rich Manipulation

Import AI 460: Reward hacking society, RSI data from Anthropic; and RL-based quadcopter racing

Shield-Loco: Shielding Locomotion Policies with Predictive Safety Filtering

Brendan Foody (Mercor) — Agentic Data and the Future of AI

FlowPRO: Reward-Free Reinforced Fine-Tuning of Flow-Matching VLAs via Proximalized Preference Optimization

GRAIL: Generating Humanoid Loco-Manipulation from 3D Assets and Video Priors

Humanoid-GPT: Scaling Data and Structure for Zero-Shot Motion Tracking

RDGen: Demonstration Generation for High-Quality Robot Learning via Reinforcement Learning

VLA-Pro: Cross-Task Procedural Memory Transfer for Vision-Language-Action Models

How to Instruct Your Robot: Dense Language Annotations Power Robot Policy Learning

BORA: Bridging Offline Reinforcement Learning and Online Residual Adaptation for Real-World Dexterous VLA Models

EXPO-FT: Sample-Efficient Reinforcement Learning Finetuning for Vision-Language-Action Models

Humanoid Everyday: A Comprehensive Robotic Dataset for Open-World Humanoid Manipulation

How Cursor Trained Composer on Fireworks: Distributed Infrastructure for High-Performance RL

When Absolute State Fails: Evaluating Proprioceptive Encodings for Robust Manipulation

Deepu Talla (NVIDIA) on the Three-Computer Physical AI Stack

Guide, Think, Act: Interactive Embodied Reasoning in Vision-Language-Action Models

166: 许华哲再次具身创业：不想错过最大的西瓜

Towards Long-horizon Embodied Agents with Tool-Aligned Vision-Language-Action Models

How to 10x any AI skill using Karpathy's Autoresearch method

Factored Diffusion Policies:Compositionally Generalized Robot Control with a Single Score Network

Judge, Then Drive: A Critic-Centric Vision Language Action Framework for Autonomous Driving

Superhuman Safe and Agile Racing through Multi-Agent Reinforcement Learning

Import AI 457: AI stuxnet; cursed Muon optimizer; and positive alignment

165: 英伟达 GEAR 高深远：世界模型、自进化循环、DreamDojo

Eric Jang – Building AlphaGo from scratch

RLDX-1 Technical Report

Energy, Minerals, and the Physical Stack Behind AI

What to Ignore, What to React: Visually Robust RL Fine-Tuning of VLA Models

Import AI 455: AI systems are about to start building themselves.

$\Delta$VLA: Prior-Guided Vision-Language-Action Models via World Knowledge Variation

Beyond Bigger Models: Recursion As The Next Scaling Law In AI

Learning while Deploying: Fleet-Scale Reinforcement Learning for Generalist Robot Policies

Jim Fan — VLAs Are Dead, WAMs Are Next

Andrej Karpathy: From Vibe Coding to Agentic Engineering

LaST-R1: Reinforcing Action via Adaptive Physical Latent Reasoning for VLA Models

163: 详解DeepSeekV4：Infra巨鲸、百万上下文走进现实、极致效率优化

Why We Need Continual Learning

OpenAI and Microsoft Make Nice

dWorldEval: Scalable Robotic Policy Evaluation via Discrete Diffusion World Model

138. 对罗福莉3.5小时访谈：AI范式已然巨变！OpenClaw、Agent范式很吃后训练、卡的分配、组织平权

Cortex 2.0: Grounding World Models in Real-World Industrial Deployment

137. 对洪乐潼的4小时访谈：AI for Math、把数学变成Lean、数学天书中的证明、直觉、被创造与被发现的

FLASH: Fast Learning via GPU-Accelerated Simulation for High-Fidelity Deformable Manipulation in Minutes

Novel Algorithms for Smoothly Differentiable and Efficiently Vectorizable Contact Manifold Construction

From Models to Mobility: Building Waymo with Dmitri Dolgov

DEX-Mouse: A Low-cost Portable and Universal Interface with Force Feedback for Data Collection of Dexterous Robotic Hands

Switch: Learning Agile Skills Switching for Humanoid Robots

How a 20-Person Startup Won Gold at the Math Olympiad—Tying With OpenAI & DeepMind (Tudor Achim, CEO of Harmonic)

Learning Versatile Humanoid Manipulation with Touch Dreaming

Simulator Adaptation for Sim-to-Real Learning of Legged Locomotion via Proprioceptive Distribution Matching

Towards Practical World Model-based Reinforcement Learning for Vision-Language-Action Models

Jensen Huang: 10 Lessons From the CEO Building the Most Important Company in History

Learning Humanoid Standing-up Control across Diverse Postures

RPL: Learning Robust Humanoid Perceptive Locomotion on Challenging Terrains

Sumo: Dynamic and Generalizable Whole-Body Loco-Manipulation

CMP: Robust Whole-Body Tracking for Loco-Manipulation via Competence Manifold Projection

Large Reward Models: Generalizable Online Robot Reward Generation with Vision-Language Models

Adaptive Action Chunking at Inference-time for Vision-Language-Action Models

Not All Features Are Created Equal: A Mechanistic Study of Vision-Language-Action Models

ABot-M0: VLA Foundation Model for Robotic Manipulation with Action Manifold Learning

Marc Andreessen on AI Winters and Agent Breakthroughs

An AI state of the union: We’ve passed the inflection point, dark factories are coming, and automation timelines | Simon Willison

ForeAct: Steering Your VLA with Efficient Visual Foresight Planning

157: 具身季报26Q1：人形再思考、英伟达世界模型、高自由度灵巧手

Functional Force-Aware Retargeting from Virtual Human Demos to Soft Robot Policies

SMASH: Mastering Scalable Whole-Body Skills for Humanoid Ping-Pong with Egocentric Vision

Sergey Levine - Building LLMs for the Physical World - [Invest Like the Best, EP.465]

Coordinated Humanoid Manipulation with Choice Policies

DIAL: Decoupling Intent and Action via Latent World Modeling for End-to-End VLA

Import AI 451: Political superintelligence; Google's society of minds, and a robot drummer

SOLE-R1: Video-Language Reasoning as the Sole Reward for On-Robot Reinforcement Learning

Scaling World Model for Hierarchical Manipulation Policies

134. 和谢晨聊“数据的综述”：AI和机器人数据的历史、版图、金字塔与Recipe

VLAW: Iterative Co-Improvement of Vision-Language-Action Policy and World Model

How François Chollet Is Building A New Path To AGI

A Unified and General Humanoid Whole-Body Controller for Versatile Locomotion

OmniRetarget: Interaction-Preserving Data Generation for Humanoid Whole-Body Loco-Manipulation and Scene Interaction

Open-World Task and Motion Planning via Vision-Language Model Generated Constraints

Steerable Vision-Language-Action Policies for Embodied Reasoning and Hierarchical Control

Why One Superintelligence Is More Dangerous Than a Thousand (Vincent Weisser, CEO & Co-Founder of Prime Intellect)

How Autonomous Labs Will Transform Scientific Research: Ginkgo Bioworks’ Jason Kelly

DualCoT-VLA: Visual-Linguistic Chain of Thought via Parallel Reasoning for Vision-Language-Action Models

Andrej Karpathy on Code Agents, AutoResearch, and the Loopy Era of AI

Scaling Sim-to-Real Reinforcement Learning for Robot VLAs with Generative 3D Worlds

OmniVTA: Visuo-Tactile World Modeling for Contact-Rich Robotic Manipulation

Why Robots Still Struggle With Simple Tasks (And What Might Finally Change That) | Karol Hausman, Co-Founder & CEO…

Bernt Børnich (1X) — Meet NEO, Your Robot Butler in Training

Jagdeep Singh (Rhoda AI) — 10 Hours, Not 10,000: $450M to Train Robots on YouTube

Pete Florence (Generalist) — Scaling Laws, Train One Improve All

Pete Florence (Generalist) — Why Robot Intelligence Will Be General

From "Large and Strong" to "Small and Strong": Density Laws, RL Scaling Laws, and the Distributed Future of Intelligence

No Priors Ep. 142 | With Harvey Co-Founder and President Gabe Pereyra

Turing CEO, Jonathan Siddharth: Who Wins in Data Labelling & Why SaaS Will Die?

Michael Levin: Reality is an Illusion - Alien Intelligence, Biology, Life | Lex Fridman Podcast #486

Interview with DeepMind's Jie Tan: Robotics, Cross-Embodiment, World Models, Gemini Robotics 1.5, and Google

Ilya Sutskever – We're moving from the age of scaling to the age of research

142: How a Self-Teaching AI Product Emerges | Dialogue with Zebra CPO Xiu Jiaming

No Priors Ep. 141 | With Sunday Robotics Co-Founders Tony Zhao and Cheng Chi

140: AI for Science, From the Beginning to Now | Dialogue with Zhang Linfeng and Sun Weijie from DeepWisdom

Amjad Masad & Adam D’Angelo: How Far Are We From AGI?

Interview with Li Xiang (Part 2): CEO Large Model, MoE, Liang Wenfeng, VLA, Energy, Memory, Confronting Human Nature, Intimate Relationships, Human Wisdom

Andrej Karpathy — “We’re summoning ghosts, not building animals”

Fully autonomous robots are much closer than you think – Sergey Levine

AI Robots for Your Laundry w/Lindon Gao LIVE FROM CLEAN - S6E103 The Laundromat Millionaire Show✈️