odesha

Welcome in 2025 - the robotic age

@StockMKTNewz I literally felt shivers when reading this!

Meta Platforms $META CEO Mark Zuckerberg has reportedly moved his desk into the AI lab working directly with Alexandr Wang and Nat Friedman Zuckerberg is reportedly coding throughout the day - Reuters

@Polymarket Oh Ohhhhh, unitree competition?

JUST IN: Alibaba reportedly preparing to unveil its first robot.

VLAs are cooked the robotics community spent 2 years building VLAs on top of language model backbones. Turns out teaching a robot physics through static image-text pairs is like learning to drive by reading the owner’s manual mimic-video just open sourced a completely different approach. video-action models (VAMs). freeze a pretrained video generation model, attach a tiny action decoder, let the video backbone handle the physics Results: 10x sample efficiency, 2x convergence speed vs VLAs. 77% success rate with ONE demo per task the insight is almost obvious in hindsight. Video models already understand how the physical world evolves over time. that’s literally what robot control needs. VLAs tried to reconstruct that understanding from scratch using robot data built on cosmos-predict2 (2B params). small enough to actually deploy. The real unlock: as video generation models improve, robot policies improve for free. No new robot data needed. VLAs had their moment. the foundation for physical AI was never language, it was video Check it out github.com/mimic-video/mi…

@lukas_m_ziegler Insane, thanks for sharing. Quite expensive still ~6k+ nearly half of it are actuators. It shows that we need more actuator manufacturing in eu and bring cost down

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! 🔥 Here's the project page: github.com/enactic/OpenArm ~~ ♻️ Join the weekly robotics newsletter, and never miss any news → ziegler.substack.com

@chris_j_paxton Fully agree!

Increasingly i think one of the biggest differentiators between who will be able to "pull off" generalized robotics intelligence is based on who understands hardware. At the margins, little details matted, and robots need to work well. So something like this -- building a humanoid robot from scratch -- seems really cool and useful

@benjitaylor @0xDesigner I don’t agree. You must love what you are doing. Doing it only for the reason can be a waste of time. Market research, customer interviews and overal validation early one is crucial. Challenge my opinion

Making something people love is mostly making something you love and hoping the overlap is real

@pvdyck1 Clean, to the point. Good luck!

@oskrt_dvs indie.money. n8n workflow marketplace. publish your automations, buyers pay per run. no monthly subs, just execution fees settled on Base. indie.money

What are you building today, Sunday 12th

@FlexAppeall You can build one! Or fill the gap, we need this in Europe. Build build build

Why can I not buy a robotic arm that gets here within a week. Why are they all from China

@elonmusk Collective knowledge winning, always

“They are smart … for humans” – future AI

@varcc888 Cool! Drop the url when it’s live!

@oskrt_dvs We are creating viberankdev, a community driven platform to explore and rank the most innovative tech tools available.

@nikitabier They trying to poach you?

Sorry guys, I’m stepping up to the big leagues now.

@BaslyAsma PTP time sync

What tool in your robotics stack do you secretly hate the most?

spent the day with an old belgian friend in zurich. runs a construction company. some things that stuck with me: europe has a real shortage of people who actually want to do manual work. young guys can be trained but the determination isn’t there. and his business isn’t making commodity products. small batch, custom pieces, every job different. you can’t automate that. not with robotics, not with ai. you need people, you need networks, you build your own systems. no playbook exists.

Well done! Netherlands has been on the fore front of electric vehicles and autonomy. From first fast charging station from FastNet and world leading research in full driving from the Tu Delft autonomous vehicle lab ME faculty. It was a long way here, and much more effort is still needed to continue deploying Europe wide

First (supervised) FSD approval in Europe! Congratulations to the Tesla team and thank you to the regulatory authorities in the Netherlands for all the hard work required to make this happen.

@elonmusk then everyone else is just stupid 🤣

Relatively smart for a human, but not by AI standards

@antoinemarcel In you are looking for contributors hmu!

@oskrt_dvs it's mine ahah, and yes we will opensource it by the end of the year

since I got a robot, I’ve been getting lazier and lazier

@Yampeleg trained on more codebases than the rest, earlier on than everyone else?

What is Anthropic’s secret that makes the Claude models the way they are?

@petergyang They are all, everywhere. Even in Switzerland! I heard Swisscom (telecom company) runs Chinese model on its own datacenters. And just sell their ai services. Basically only paying inference. Absolutely crazy

Silicon Valley is quietly running on Chinese open source AI models. Here are the receipts: → Cursor confirmed last month that Composer 2 is built on Moonshot's Kimi K2.5 → Cognition's SWE-1.6 model is likely post-trained on Zhipu's GLM → Shopify saved $5M a year by switching to Alibaba’s Qwen model. Airbnb CEO Brian Chesky has also said: "We rely a lot on Qwen. It's very good, fast, and cheap." And now Zhipu dropped GLM-5.1, an open source model that performs almost as well as Opus on coding benchmarks. 📌 More on the Anthropic + OpenClaw drama and what I'm learning about AI on the ground in China in my new post: creatoreconomy.so/p/the-all-you-…