Zu Wang
@zuwang95
Followers
153
Following
653
Statuses
452
Intern at NVIDIA GEAR, NYU GRAIL. 25' Undergrad at NYU. Previously SJTU 23'. Graphics, Robotics. Real Fantasy.
New York, NY | Santa Clara, CA
Joined December 2022
We need to establish a scalable societal framework for real-world data collection, especially in robotics. Consider autonomous driving as an example: cars inherently provide value to individuals, making widespread deployment feasible. This utility naturally facilitates data collection, as users contribute by simply using their vehicles. A similar strategy should be adopted for robots. If robots can perform valuable tasks and are deployed in millions of households, people would likely engage in teleoperating their robots to teach them new skills. This could initiate a "robot data flywheel," where the utility of robots drives data collection, and the resulting data, in turn, enhances their capabilities, creating a self-reinforcing cycle. This is why works like UMI and RUM stand out—they offer a practical blueprint for the future of robot data collection. Ultimately, we cannot rely solely on financial incentives to drive data collection.
my math factored in that you might just rent 1000-10000 different apartments, each with 1-10 robots working 24/7. Apartment rental costs can get up to 10,000,000 range easily. The time it takes to set the apartment up for all kinds of “lived in” scenarios (stuff on floor, plants, dirty dishes) adds more cost per apartment + time use
2
5
19
RT @TairanHe99: This fellowship is a recognition of joint efforts of all my collaborators! Thank you!!!!
0
4
0
RT @CMU_Robotics: Congratulations to Ph.D. @TairanHe99, who received the '25-'26 @nvidia Graduate Fellowship! 👏 Tairan received the award…
0
4
0
RT @zhengyiluo: I remember first seeing this and thought “oh snap MPCs? Should we go back to classic control for locomotion & manipulation?…
0
1
0
RT @ToruO_O: When @zhaohengyin first told me about this idea, I wasn't impressed. "How far can you get with teleoperation for hands?" I ask…
0
3
0
RT @Vikashplus: @KarlPertsch @LerrelPinto @notmahi @gupta_abhinav_ @svlevine @Ken_Goldberg @haoshu_fang Kudos @LerrelPinto for opening this…
0
2
0
RT @LerrelPinto: @KarlPertsch Hi Karl, what do you mean by a 'generalist' manipulation policy? If you mean a policy that can do a standar…
0
2
0
Pi open source, huge win for the whole robotics community!
Many of you asked for code & weights for π₀, we are happy to announce that we are releasing π₀ and pre-trained checkpoints in our new openpi repository! We tested the model on a few public robots, and we include code for you to fine-tune it yourself.
0
0
2
RT @_wenlixiao: 🚨Apparently, Sim2Real is NOT ENOUGH to solve humanoid whole-body control 🎉To push humanoid agility to the next level, we de…
0
11
0
👑
🚀 Can we make a humanoid move like Cristiano Ronaldo, LeBron James and Kobe Byrant? YES! 🤖 Introducing ASAP: Aligning Simulation and Real-World Physics for Learning Agile Humanoid Whole-Body Skills Website: Code:
0
0
1
RT @SongShuran: 🚀 Meet ToddlerBot 🤖– the adorable, low-cost, open-source humanoid anyone can build, use, and repair! We’re making everythi…
0
18
0
RT @HaochenShi74: Time to democratize humanoid robots! Introducing ToddlerBot, a low-cost ($6K), open-source humanoid for robotics and AI r…
0
98
0
I immediately shared this with my lab mates, all of us are deeply impressed by the design and capability of this cute humanoid! Really amazing work, congrats!
Checkout how we spent our last year! Now everyone can build their own humanoid and start research with just a 3D printer and some basic tools!
1
0
2
RT @GaoyueZhou: Can we extend the power of world models beyond just online model-based learning? Absolutely! We believe the true potential…
0
99
0
RT @ericjang11: Why is the market selling off Nvidia / compute stocks? The correct reaction to R1 breakthrough should be to buy *even more*…
0
94
0
💯
Today's "DeepSeek selloff" in the stock market -- attributed to DeepSeek V3/R1 disrupting the tech ecosystem -- is another sign that the application layer is a great place to be. The foundation model layer being hyper-competitive is great for people building applications.
0
0
0