Tony Tao (@_tonytao

Tony Tao

308 posts

Tony Tao

@_tonytao_

Incoming PhD @berkeley_ai | Prev @CMU_Robotics

Joined June 2024

Pinned
Tony Tao
@_tonytao_
May 13, 2025
Training robots for the open world needs diverse data But collecting robot demos in the wild is hard! Presenting DexWild 🙌🏕️ Human data collection system that works in diverse environments, without robots 💪🦾 Human + Robot Cotraining pipeline that unlocks generalization 🧵👇
00:00
97K
Tony Tao
@_tonytao_
Aug 20, 2025
this is possibly the best whole body teleoperation that I've ever seen. The motion is really smooth and dynamic. The autonomous results are equally impressive. bostondynamics.com/blog/large-beh…
00:00
58K
Tony Tao
@_tonytao_
Aug 6, 2025
sometimes it feels like we are just going in circles in robotics
ARX
@ARXrobotics
Aug 6, 2025
Introducing AC one: Data collection efficiency increased by up to 10 times.🦾 Allows first-time users to get started in no time.🎆 Delivers impressive results on flexible objects.🧶 More details will be released in the coming weeks.
00:00
30K
Tony Tao
@_tonytao_
Jun 21, 2025
Thrilled to have received Best Paper Award at the EgoAct Workshop at RSS 2025! 🏆 We’ll also be giving a talk at the Imitation Learning Session I tomorrow, 5:30–6:30pm. Come to learn about DexWild! Work co-led by @mohansrirama, with @JasonJZLiu, @kenny__shaw, and @pathak2206.
20K
Tony Tao
@_tonytao_
Jun 17, 2025
🦾 DexWild is now open-source! Scaling up in-the-wild data will take a community effort, so let’s work together. Can’t wait to see what you do with DexWild! Main Repo: github.com/dexwild/dexwild Hardware Guide: tinyurl.com/dexwild-hardwa… Training Code: github.com/dexwild/dexwil…
Tony Tao
@_tonytao_
May 13, 2025
Training robots for the open world needs diverse data But collecting robot demos in the wild is hard! Presenting DexWild 🙌🏕️ Human data collection system that works in diverse environments, without robots 💪🦾 Human + Robot Cotraining pipeline that unlocks generalization 🧵👇
00:00
GitHub - dexwild/dexwild: DexWild: Dexterous Human Interactions for In-the-Wild Robot Policies
From github.com
9.9K
Tony Tao
@_tonytao_
Jul 2, 2025
Whenever I see something impressive in robotics, I always ask for a live demo. If they can’t demo it, it’s hard to believe it really works. I hope more companies do demos in the future to show the world their policies work beyond a cool video.
Dyna Robotics
@DynaRobotics
Jul 2, 2025
We have started taking DYNA-1, our dexterous robust VLA model, to conferences and showcasing it for hours on end! The model run for 3 days, 8 hours each day at #HITEC2025 3 weeks ago with 99.9% overall success rate (dropped 1 towel in day 2). No intervention, it just works :)
00:00
6.1K
Tony Tao
@_tonytao_
Aug 7, 2025
Vision based locomotion ftw!!!
Skild AI
@SkildAI
Aug 6, 2025
We’ve all seen humanoid robots doing backflips and dance routines for years. But if you ask them to climb a few stairs in the real world, they stumble! We took our robot on a walk around town to environments that it hadn’t seen before. Here’s how it works🧵⬇️
00:00
4.5K
Tony Tao
@_tonytao_
Jul 9, 2025
Want to add diverse, high-quality data to your robot policy? Happy to share that the DexWild Dataset is now fully public, hosted by @huggingface 🤗 Find it here! huggingface.co/datasets/board…
01:06
Tony Tao
@_tonytao_
May 13, 2025
Training robots for the open world needs diverse data But collecting robot demos in the wild is hard! Presenting DexWild 🙌🏕️ Human data collection system that works in diverse environments, without robots 💪🦾 Human + Robot Cotraining pipeline that unlocks generalization 🧵👇
5.9K
Tony Tao
@_tonytao_
Jun 24, 2025
Dexwild Live Demo today at #RSS2025! Come by our table at Associates Park to see us zero shot our policy in-the-wild 🌿
00:00
2.8K
Tony Tao
@_tonytao_
Jul 21, 2025
Had a lot of fun chatting with @chris_j_paxton and @micoolcho! In the episode, we unpack the hidden stories, interesting detours, and lessons learned behind our paper DexWild 🦾
RoboPapers
@RoboPapers
Jul 21, 2025
Full episode dropping soon! Geeking out with @_tonytao_ @mohansrirama on DexWild - Dexterous Human Interactions for In-the-Wild Robot Policies dexwild.github.io Co-hosted by @chris_j_paxton @micoolcho
00:00
3.1K
Tony Tao
@_tonytao_
Jun 10, 2025
we had problems with temporal ensembling using diffusion policy due to these exact inference delay and multimodality issues a few months ago, having to resort to rolling out actions open loop. really excited to try this out!
Sergey Levine
@svlevine
Jun 9, 2025
Real-time inference is a big challenge for VLAs. We’ve been working on a way to amortize inference delays in π0.5. Our new Real-Time Chunking (RTC) method speeds up π0.5 by allowing the robot to “think” while it’s moving, which makes it quite a bit faster! 🧵👇
00:00
2.2K
Tony Tao
@_tonytao_
May 17, 2025
I'll be at ICRA next week. Hoping to meet some new friends! 😄
1.6K
Tony Tao
@_tonytao_
Aug 8, 2025
this is such a cool idea! wonder if we can do something similar for robotics.
Lili
@lchen915
Aug 8, 2025
Self-Questioning Language Models: LLMs that learn to generate their own questions and answers via asymmetric self-play RL. There is no external training data – the only input is a single prompt specifying the topic.
3.3K
Tony Tao
@_tonytao_
May 13, 2025
Replying to @_tonytao_
Diversity and Quality are key to our success! We collect 9,290 human demos across 93 environments — at 5× the speed of teleop methods. A team of untrained collectors gathers data across objects, environments, and tasks to get the diversity we need! 2/n
00:00
2.2K