Even this MD skill per device seems a bit insane and just like burning tokens to me, you are taking something deterministic and making it non-deterministic
github.com/facebookincubator…
But I guess this is the future, just not the one I’ll take part in sorry seems just wasteful
The future of world models will not be about scaling compute.
It will be about scaling structure.
Mechanistic models aim to learn the underlying dynamics of the world rather than brute-force intelligence with ever-larger models and more compute.
This is why we’re betting on mechanistic world models.
analyticsindiamag.com/ai-new…
Humanoids shouldn't need a nine-figure budget.
Meet CoPAEI: an open, modular humanoid platform, made in the USA.
Founding Cohort now open: 10 institutions, 2 robots each.
Request access: copaei.usvz.ai
Grateful to @CharacterCap@sachgaur for backing us early.
We built a humanoid robot platform called CoPAEI, a physics-aware, modular, and easy-to-repair humanoid designed to help scale humanoid robotics across America.
Building a robot is more than just writing code.
From hardware setup to software integration and testing, every stage brings its own challenges.
What's the biggest bottleneck in your robotics development workflow?
#Robotics#RobotDevelopment#PhysicalAI
We turned an AI agent loose on a college football controversy: camera calibration from goalpost geometry, frame-by-frame ball tracking, and a drag physics fit, all from fan videos. Forensic sports physics in an afternoon. @grok@xai@cursor_ai
JBR-001: an open-source Desktop Companion Robot powered by the Arduino® UNO™ Q and built to make Robotics, Computer Vision, Edge AI, and Physical AI fun to explore #DYI#Roboticsprojecthub.arduino.cc/synthe…
1/14 Imagine giving a robot the exact same movement twice. The first command says “move to position 0.6.” The second says “move +0.1 from where you are.” Same trajectory, same information. Yet its world model may imagine two different futures.
Your fraud detection model works perfectly in testing.
Weeks after deployment, fraud patterns change. False positives increase, and the model starts missing suspicious transactions.
The API is still running. But is the AI still performing well?
Two days ago I had 11 followers. This post changed that.
I'm now past 200, and I've met so many people building with AI. Thank you to everyone who connected.
If we haven't met yet: I build open-source tools for coding agents. What are you building? Let's #connectx.lingyaoai.com/eandualem/status/21061…
Looking to connect with more people building in these spaces 👇
- Founders
- Automation
- AI Agents
- ML Engineers
- SaaS
- Full-stack
- DevOps
- App Development
- AI / ML
- Data Science
- Freelancing
- Building in public
If that’s you, let’s #connect Drop 🤝 and let's connect.