Embodied AI 101
Episodes
Episode 61: DexWM: World Models for Dexterous Manipulation from Human Videos
20 Dec 2025
Contributed by Lukas
# DexWM: World Models for Dexterous Manipulation from Human Videos Dexterous manipulation – the art of using multi-fingered hands to pick, place, t...
Episode 60: Video-Action Models: Generalizing Robot Control with Video Diffusion
19 Dec 2025
Contributed by Lukas
# Video-Action Models: Generalizing Robot Control with Video Diffusion In a new preprint titled *“mimic-video: Video-Action Models for Generalizabl...
Episode 59: Calibrated Confidence in Controllable Video Models
19 Dec 2025
Contributed by Lukas
# Calibrated Confidence in Controllable Video Models In the recent preprint *“World Models That Know When They Don't Know: Controllable Video Gener...
Episode 58: DexWM: World-Modeling Dexterous Manipulation from Human Videos
19 Dec 2025
Contributed by Lukas
# DexWM: World-Modeling Dexterous Manipulation from Human Videos In this episode we dive into a new framework for dexterous manipulation: **“World ...
Episode 57: Scaling Up Offline Model-Based RL with Action Chunks (MAC)
19 Dec 2025
Contributed by Lukas
# Scaling Up Offline Model-Based RL with Action Chunks (MAC) In *“Scalable Offline Model-Based RL with Action Chunks”* (Park et al., 2025) – by...
Episode 56: Emergent Human-to-Robot Transfer in Vision-Language-Action Models
19 Dec 2025
Contributed by Lukas
# Emergent Human-to-Robot Transfer in Vision-Language-Action Models **Simar Kareer, Karl Pertsch, James Darpinian, Judy Hoffman, Danfei Xu, Sergey Le...
Episode 55: DexScrew: Learning Dexterous Manipulation from “Imperfect” Simulations
15 Dec 2025
Contributed by Lukas
# DexScrew: Learning Dexterous Manipulation from “Imperfect” Simulations *Learning Dexterous Manipulation Skills from Imperfect Simulations* is a...
Episode 54: X-Humanoid: Robotizing Human Videos into Humanoid Videos
15 Dec 2025
Contributed by Lukas
# X-Humanoid: Robotizing Human Videos into Humanoid Videos A new preprint titled **“X-Humanoid: Robotize Human Videos to Generate Humanoid Videos a...
Episode 53: Decoupled Q-Chunking: Combining Long-Range Value Propagation with Reactive Policies
13 Dec 2025
Contributed by Lukas
# Decoupled Q-Chunking: Combining Long-Range Value Propagation with Reactive Policies In *Decoupled Q-Chunking* (Li, Park, Levine, 2025), the authors...
Episode 52: F2D2: Joint Distillation for Fast Likelihood and Sampling in Flow Models
11 Dec 2025
Contributed by Lukas
# F2D2: Joint Distillation for Fast Likelihood and Sampling in Flow Models In the recent preprint **“Joint Distillation for Fast Likelihood Evalu...
Episode 51: Training-Time Action Conditioning for Efficient Real-Time Chunking
09 Dec 2025
Contributed by Lukas
# Training-Time Action Conditioning for Efficient Real-Time Chunking In “Training-Time Action Conditioning for Efficient Real-Time Chunking” (Bla...
Episode 50: π0.6: Learning from Experience for Vision–Language–Action Robotic Models
07 Dec 2025
Contributed by Lukas
In a recent technical report, the Physical Intelligence team (led by Ali Amin, Raichelle Aniceto, Ashwin Balakrishna, Kevin Black, Ken Conley, Grace C...
Episode 49: Robotic World Model (RWM): Learning Stable Neural Simulators for Long-Horizon Control
07 Dec 2025
Contributed by Lukas
# Robotic World Model (RWM): Learning Stable Neural Simulators for Long-Horizon Control Recent work by Chenhao Li, Andreas Krause, and Marco Hutter (...
Episode 48: Much Ado About Noising: Dispelling the Myths of Generative Robotic Control
06 Dec 2025
Contributed by Lukas
# Much Ado About Noising: Dispelling the Myths of Generative Robotic Control In recent years, robotics and control researchers have widely embraced *...
Episode 47: ReWiND: Language-Guided Reward Learning for Robot Task Adaptation
06 Dec 2025
Contributed by Lukas
# ReWiND: Language-Guided Reward Learning for Robot Task Adaptation Modern robot learning systems crave rich supervision, but collecting demonstratio...
Episode 46: Nested Learning: Unifying Architecture and Optimization for Memoryful Models
05 Dec 2025
Contributed by Lukas
# Nested Learning: Unifying Architecture and Optimization for Memoryful Models In “Nested Learning: The Illusion of Deep Learning Architectures” ...
Episode 45: AsyncVLA: Real-Time Vision-Language-Action through Asynchronous Flow Matching
04 Dec 2025
Contributed by Lukas
# AsyncVLA: Real-Time Vision-Language-Action through Asynchronous Flow Matching **Context & Motivation.** Robotics is riding a wave of foundation mo...
Episode 44: GigaWorld-0: World Models as a Data Engine for Embodied AI
04 Dec 2025
Contributed by Lukas
# GigaWorld-0: World Models as a Data Engine for Embodied AI **1. What Problem Is This Paper Trying to Solve?** Modern embodied AI — controlling r...
Episode 43: AAWR: Using Privileged Training to Learn Active Perception in Robotics
03 Dec 2025
Contributed by Lukas
# AAWR: Using Privileged Training to Learn Active Perception in Robotics Modern robots often must act under partial observability: their onboard sens...
Episode 42: GR-RL: Turning a Generalist VLA Policy into a Dexterous Dexterous Manipulator
03 Dec 2025
Contributed by Lukas
# GR-RL: Turning a Generalist VLA Policy into a Dexterous Dexterous Manipulator Modern robotics research is increasingly centered on large, general...
Episode 41: ReWiND in context: how language-guided rewards compare to 2024-2025 alternatives
02 Dec 2025
Contributed by Lukas
**ReWiND represents a significant contribution to language-conditioned reward learning, achieving 2x performance improvements over baselines through n...
Episode 40: Stage-Aware Reward Modeling (SARM) for Long-Horizon Manipulation
02 Dec 2025
Contributed by Lukas
# Stage-Aware Reward Modeling (SARM) for Long-Horizon Manipulation **Summary of the source** – Chen *et al.* (Sept. 2025) introduce *SARM*, a **...
Episode 39: RL-100: Performant Robotic Manipulation with Real-World Reinforcement Learning
02 Dec 2025
Contributed by Lukas
# RL-100: Performant Robotic Manipulation with Real-World Reinforcement Learning **Main claim:** RL-100 introduces a three-stage learning pipeline th...
Episode 38: ResMimic: Two-Stage Residual Learning for Humanoid Loco-Manipulation
01 Dec 2025
Contributed by Lukas
# ResMimic: Two-Stage Residual Learning for Humanoid Loco-Manipulation ResMimic addresses the challenging problem of **humanoid whole-body loco-manip...
Episode 37: HO-Cap Dataset and Capture System
01 Dec 2025
Contributed by Lukas
# HO-Cap Dataset and Capture System **HO-Cap** (Hand-Object Capture) is a newly released dataset and capture setup for hand-object interaction, intro...
Episode 36: VIRAL rewrites the sim-to-real playbook for humanoid robots
27 Nov 2025
Contributed by Lukas
The path from simulated robot to real-world deployment has long been paved with domain randomization and crossed fingers. VIRAL, a new paper from NVID...
Episode 35: ENACT: A New Benchmark for Embodied Cognition in Home Robotics
26 Nov 2025
Contributed by Lukas
**Why ENACT matters to you:** Traditional AI benchmarks often use static images or short navigation tasks, but home robots live in a **dynamic, intera...
Episode 34: In-N-On: Scaling Egocentric Manipulation with In-the-Wild and On-Task Data
24 Nov 2025
Contributed by Lukas
**Introduction.** Foundation models for robot manipulation have begun leveraging **egocentric human demonstration data** as a rich resource for learni...
Episode 33: How Meta Trained SAM‑3D: Extending “Segment Anything” to 3D
20 Nov 2025
Contributed by Lukas
**Meta’s SAM‑3D** is a new **foundation model for 3D** vision that extends the 2D **Segment Anything Model (SAM)** into the third dimension. It ac...
Episode 32: Closing the Sim-to-Real Loop: GSWorld’s Photorealistic Real2Sim2Real Pipeline
18 Nov 2025
Contributed by Lukas
## Motivation: Why Sim-to-Real Matters Now Robotic manipulation and navigation systems are reaching levels of complexity that make **sim-to-real te...
Episode 31: WMPO: Training Robot Policies in Imagination
17 Nov 2025
Contributed by Lukas
## Introduction – From Imitation to Imagination Imagine a household robot that can **follow human instructions** flawlessly when everything goes ...
Episode 30: Learning a Thousand Tasks: Oxford’s MT3 and the Quest for Generalist Robots
14 Nov 2025
Contributed by Lukas
## Introduction Robots that can learn **many diverse tasks** efficiently have long been a goal in robotics. Humans can watch a single demonstration...
Episode 29: BFM-Zero: Unsupervised RL for Humanoid Control with a Shared Latent Skill Space
12 Nov 2025
Contributed by Lukas
## Introduction Building generalist robot controllers that can perform many tasks on demand is a major goal in vision-language-action (VLA) models ...
Episode 28: Self Forcing: Bridging the Train–Test Gap in Autoregressive Video Diffusion
09 Nov 2025
Contributed by Lukas
Recent advances in text-to-video generation have achieved impressive fidelity and complex temporal dynamics in short clips. However, many state-of-the...
Episode 27: Smol-Training Playbook: Technical Review of Tools and Frameworks
08 Nov 2025
Contributed by Lukas
## Introduction The **Smol-Training Playbook** from Hugging Face is a comprehensive guide that distills best practices for training large language ...
Episode 26: SoftMimic: Teaching Humanoids to Move Softly
07 Nov 2025
Contributed by Lukas
## The Problem with Rigid Controllers Most robots trained to imitate human motions have a hidden flaw: they’re **too stiff**. When an imitation-tra...
Episode 25: GEN‑0: Pushing the Frontiers of Embodied AI with Harmonic Reasoning
07 Nov 2025
Contributed by Lukas
Modern embodied AI models are rapidly evolving, and _GEN‑0_ by Generalist AI represents a significant leap. Announced in late 2025, GEN‑0 is an em...
Episode 24: DINOv3 and the Next Generation of Visual Foundation Models
15 Aug 2025
Contributed by Lukas
Hello and welcome to Embodied AI 101. Today, we dive into a critical review of DINOv3, a 2025 vision model from Meta AI that marks a major step to...
Episode 23: A Critical Look at Hume VLA
12 Aug 2025
Contributed by Lukas
Introduction – Two Minds Inside One Robot Modern embodied AI is drawing inspiration from human cognition. Psychologist Daniel Kahneman famous...
Episode 22: Critical Review of π0.5
11 Aug 2025
Contributed by Lukas
Introduction Roboticists have long dreamed of generalist robots that can step out of the lab and perform useful tasks in unstructured, everyday...
Episode 21: Deep Dive: ReinboT and the Fusion of RL with Vision-Language-Action
04 Aug 2025
Contributed by Lukas
Introduction: A New Twist in Robot Learning Hello and welcome to Robotics Unwrapped, where we explore cutting-edge advances in robot learning. ...
Episode 20: RIPT-VLA - The Fine-Tuning Revolution
01 Aug 2025
Contributed by Lukas
Section 1: The Distributional Shift Problem At the heart of modern robotics lies a fundamental challenge. We train our most advanced models, kn...
Episode 19: The Key to Adaptable Robots: Reinforcement Learning
01 Aug 2025
Contributed by Lukas
Imagine a robot helper in your home. You ask it, "Hey, could you put the milk in the fridge?" Simple enough. But what if you bought a different br...
Episode 18: A Technical Blueprint for Your Own Sim2Real Project
31 Jul 2025
Contributed by Lukas
In our last episode, we saw how foundation models provide robots with a "common sense" understanding of the world. But a passive understanding is ...
Episode 17: The Role of Foundation Models in Sim2Real
31 Jul 2025
Contributed by Lukas
Welcome to the final episode of our series on the Sim2Real challenge. We've been on a long journey, from dissecting the "reality gap" to exploring...
Episode 16: Transfer Learning and Meta-Learning in Sim2Real
31 Jul 2025
Contributed by Lukas
Welcome back. In our previous episodes, we've explored powerful techniques like Domain Randomization and Adversarial Learning. These methods are a...
Episode 15: Adversarial Approaches to Sim2Real
31 Jul 2025
Contributed by Lukas
Welcome back to our series on the Sim2Real challenge. In our last episode, we explored Domain Randomization, a technique where we embrace chaos in...
Episode 14: Domain Randomization: A Key Technique for Sim2Real Transfer
31 Jul 2025
Contributed by Lukas
Welcome back to the podcast. In our last episode, we dissected the "reality gap"—the chasm between the clean, predictable world of simulation an...
Episode 13: The Sim2Real Challenge: Why Virtual Robots Struggle in the Real World
31 Jul 2025
Contributed by Lukas
Welcome to the podcast, where we explore the cutting edge of AI and robotics. Today, we're diving into one of the most fundamental challenges in r...
Sim2Real Challenge
28 Jul 2025
Contributed by Lukas
The transfer of policies from simulation to physical hardware, a process known as Sim2Real, represents one of the most significant and persistent ...
Episode 5: Beyond OpenVLA – The Evolving Landscape of Vision-Language-Action Systems
23 Jul 2025
Contributed by Lukas
This is it – our final episode in the series. So far, we’ve focused on OpenVLA itself. Now we’re zooming out to the bigger picture of Vision...
Episode 4: From Simulation to Reality – Embodiment and Real-World Deployment
23 Jul 2025
Contributed by Lukas
Welcome back! So far we’ve tackled the concept of OpenVLA and the training of OpenVLA. Now it’s time for the real fun: robots! In this episode...
Episode 3: Training a Robot’s Brain – OpenVLA’s Learning and Adaptation
23 Jul 2025
Contributed by Lukas
Welcome back to our OpenVLA deep dive. In the last episode we figured out what the model’s parts are and how they operate. Now it’s time for t...
Episode 2: Under the Hood of OpenVLA – Architecture and Inference
23 Jul 2025
Contributed by Lukas
Welcome back! Last time we talked about what OpenVLA is at a high level. Now it’s time to lift the hood and see how this engine runs. How can on...
Episode 1: From Vision and Language to Action – An Introduction to VLAs and OpenVLA
23 Jul 2025
Contributed by Lukas
Hello and welcome! In this first episode, we’re laying the groundwork for our journey into Vision-Language-Action systems. Today we’ll answer:...
Episode 6: The Road Ahead – GR00T N1.5 and the Future of Humanoid AI
22 Jul 2025
Contributed by Lukas
Hello and welcome to the final episode of our deep dive on NVIDIA’s GR00T N1. It’s been a fascinating journey so far, and now it’s time to l...
Episode 5: Real Robots, Real Results – GR00T N1 in Action
22 Jul 2025
Contributed by Lukas
Welcome to Episode 5! Now that we know what GR00T N1 is capable of in theory and controlled tests, let’s explore how it’s being used in practi...
Episode 4: Skills and Smarts – What GR00T N1 Can Do
22 Jul 2025
Contributed by Lukas
Welcome back to our GR00T N1 deep dive. So far, we’ve covered the “what” and the “how” – what GR00T N1 is made of and how it was train...
Episode 3: Training a Generalist – How GR00T N1 Learned to Act
22 Jul 2025
Contributed by Lukas
Welcome back! Now that we’ve uncovered the clever architecture of NVIDIA’s GR00T N1, it’s time to answer a big question: How do you teach a ...
Episode 2: Inside GR00T N1’s Dual-System Brain
22 Jul 2025
Contributed by Lukas
Welcome back to our deep dive on NVIDIA’s GR00T N1. In the last episode, we talked about how this model is ushering in a new era of generalist r...
Episode 1: Introducing GR00T N1 – A New Era of Generalist Robots
22 Jul 2025
Contributed by Lukas
Hello and welcome! In this series, we’re diving deep into NVIDIA’s GR00T N1 model – a groundbreaking development that signals a new era for ...