@tensorfleetnet
Account based inCambodia!
About this account
- Account based in
- Cambodia
- Connected via
- United States App Store
X says this location may be affected by a proxy or VPN.
Account-level information from X, not a live location or the device used for a specific post.
Manage, coordinate, and deploy drone & robot fleets. TensorFleet gives you fleet APIs, mission graphs and ROS/PX4 adapters to go from single drone to production
Joined October 2025
- Tweets43
- Following31
- Followers48
- Likes161
This could make prototyping missions and autonomy stacks in diverse setups way smoother.
Meet SceneSmith: An agentic system that generates entire simulation-ready environments from a single text prompt.
VLM agents collaborate to build scenes with dozens of objects per room, articulated furniture, and full physics properties.
We believe environment generation is no longer the bottleneck for scalable robot training and evaluation in simulation.
Website: scenesmith.github.io/
👇🧵(1/8)
<1hr fine-tuning for that level of adaptability is a big step forward for real-world deployment testing.
The DeepMind-powered Atlas upgrade and LG’s CLOiD laundry bot show how AI models are finally turning impressive hardware into something practical. Still early, but the gap between demo and deployment feels narrower this year!
Depth Anything 3 is going to unlock a lot of interesting use cases for drones. Imagine a security drone mapping out your entire building then easily flying and checking all the floors, no longer need security guards. We can’t wait to integrate these models !
After a year of team work, we're thrilled to introduce Depth Anything 3 (DA3)! 🚀
Aiming for human-like spatial perception, DA3 extends monocular depth estimation to any-view scenarios, including single images, multi-view images, and video.
In pursuit of minimal modeling, DA3 reveals two key insights:
💎 A plain transformer (e.g., vanilla DINO) is enough. No specialized architecture.
✨ A single depth-ray representation is enough. No complex 3D tasks.
Three series of models have been released: the main DA3 series, a monocular metric estimation series, and a monocular depth estimation series.
The core team members, aside from me: @HaotongLin, Sili Chen, Jun Hao Liew, @donydchen.
👇(1/n)
#DepthAnything3
TensorFleet retweeted
K Scale Labs just shut down their open source humanoid robot project.
This is actually pointing us in the RIGHT direction.The future of open robotics isn't open source hardware—it's open source software. Here's why 🧵
How will USA compete with China in robotics?
@stash_pomichter and his team have a huge answer.
Rip out their OS and put an American-made OS into the Chinese robots.
Let's be honest. America is not going to catch up to Chinese robotics companies. The Chinese ecosystem is far more advanced than even in San Francisco. I've been visiting as many in the San Francisco ecosystem as I can and they all tell me this is true, with the exception of Tesla.
So, the world is about to be flooded with Chinese robots of all types.
Do business people anywhere, whether they are in Europe, India, America, or Asia, want to gift all their data to the Chinese?
No.
@dimensionalos has the answer, and here I visit with Stash and his team who already have an OS running on a variety of Chinese robotics.
And they are building a new kind of OS (I call it a World Model, but it doesn't matter what you call it, we are headed toward an AGI that can make a robot get smarter and smarter over time and do more and more things autonomously over time) and here Stash goes into his work and gives you a very intimate look into one of San Francisco's most interesting robotics startups I've met.
Launched our new website. Our VsCode plugin for #ROS2 is coming soon tensorfleet.net
Looks interesting we should integrate support into our VsCode extension
Build a humanoid robot yourself! 🪚
OpenArm is an open-source humanoid robot.
It comes with full CAD, control code, firmware, and simulation tools, everything needed to build, modify, and operate it.
The arms are designed to be compliant and backdrivable. Teleoperation is supported, with force feedback and real-time gravity compensation so operators can guide the arm naturally.
Super important, on the simulation side, OpenArm works with platforms like MuJoCo and Isaac Sim, letting developers test policies in virtual environments before running on hardware.
Assemble it yourself from a kit or get it prebuilt, the goal is accessibility for research labs, small teams, and enthusiasts.
The project is run by Enactic in Tokyo, Japan, and aims to lower the barrier for experimenting with dexterous manipulation.
Let's put robotics in mainstream! 🔥🔥
P.S. I put the project page in the comments!
Exciting working, we want to take some of these cool ideas into our next course
A comprehensive, hands-on tutorial on the most recent advancements in robotics 🤟
...with self-contained explanations of modern techniques for end-to-end robot learning & ready-to-use code examples using @LeRobotHF and @huggingface. Now available everywhere! 🤗
Launching a pure Python library for controlling drones. First version implemented gazebo over ZeroMq and Zenoh. Will expand to mavlink for px4 and ardupilot soon github.com/TensorFleet/easy-…
We made a fun promo in #sora2 if you want a Sora invite hit us up with have a few invite codes to give out