The bottleneck for most people who want to build a robot isn't motivation
it's not knowing where to start
what parts. what order. what software stack.
Tnkr solves exactly that — open source robot projects with step-by-step assembly, CAD, firmware, everything
https://t.co/Q2fS0D2lyU
The thing that keeps me building in this space, even on the hard days, is a belief that sounds grandiose until you actually sit with it.
We might be the generation that takes intelligence off Earth.
Not biological intelligence. We are fragile. We need air, water, a narrow band of temperature, constant supply chains, but embodied artificial intelligence can survive where we cannot. The robots being built right now are the first that could genuinely make civilization multiplanetary.
Expanding intelligence and expanding into the universe are the same project. A civilization that masters embodied AI does not stay on one rock and the work happening in labs and garages right now, the manipulation models, the world models, the open hardware, is the unglamorous foundation of exactly that.
Most people see robots folding laundry. I see the first step off the planet.
Something drops friday.
One versionable single source of truth for your entire robot.
For open source. for enterprise. for every builder.
Where the world builds robots together.
https://t.co/hkMJIL9jLt
The most important thing happening in robotics is not the biggest humanoid or the flashiest demo. It is the quiet construction of open foundations that every other builder gets to stand on.
World models. Shared datasets. Demonstration data. Simulation environments. Open VLA models you can fine-tune on a workstation. Open hardware you can 3D print and build yourself.
This is the Linux moment for robotics, and it is not a prediction. It is happening release by release, in public, right now.
A capable manipulation model that needed proprietary infrastructure and serious money two years ago can now be forked and fine-tuned by a student with a mid-range computer and public data. The gates are coming down and history says the same thing every time the gates come down. The open ecosystem out-iterates the closed one. It happened with the internet. It is happening here.
@DynamicWebPaige@Stanford@GoogleAIStudio Pupper already has a full home on Tnkr
Docs, parts list, build instructions all in one place for anyone who wants to build this themselves.
https://t.co/nQn72XKb0A
POV: you built a robot dog in your bedroom. 🐶🤖
Meet Pupper, from @Stanford! We hooked it up to @GoogleAIStudio's Gemini Live API and the Gemini Robotics-ER model so it can actually understand and react to the world.
And you can even build it yourself!
🍓 Runs on a @Raspberry_Pi
🛠️ 100% open-source (hw & sw)
We release Wuji MJLab, an open-source MuJoCo environment for dexterous hand manipulation.
It includes a cube reorientation task, a sim2real pipeline, and the setup needed to reproduce the system.
We will be at ICRA booth 121 with a live demo and welcome discussions on dexterous manipulation.
Code: https://t.co/EpJMcozjnV
Contributors: Jielin Wu, @yaoshzh, @BergerHunger, Li Chengmeng.
This project is based on @kevin_zakka’s mjlab project.
I keep getting asked where builders should document their robots.
Not GitHub. Not Notion. Not Discord.
One place. Everything connected.
@tnkrdotai 2.0 dropping soon.
Read something today that reframed how I think about robot hands.
I’ve seen plenty of humanoid companies talk about degrees of freedom, but I hadn’t spent much time thinking about the difference between degrees of freedom and actuated degrees of freedom.
The distinction is simple:
Just because a joint can move doesn’t mean the robot can actively control it.
What stood out to me about 1X’s NEO is the focus on being able to actuate and control every available degree of freedom in the hand.
It feels like one of those small details that ends up mattering a lot.
The gap between a robot that can move its fingers and a robot that can reliably manipulate objects in the real world is still massive.
Makes me think that some of the biggest breakthroughs in humanoids over the next few months to years won’t come from walking.
They’ll come from dexterity.
https://t.co/uIHEzlmlgz
Robotics today looks like software before GitHub.
Brilliant people. Building in isolation. Repeating the same work. Knowledge walks out the door every time
someone leaves.
Code on GitHub. CAD on Onshape. Docs on Notion. Wiring in Discord. Nothing in sync.
@tnkrdotai is fixing that. Friday.
@NileshArnaiya Appreciate that
All of this lives on https://t.co/ZixW9yFOJm now come join the builder community on Discord too, we'd love to have you https://t.co/bIEeV0HVVC