- Superhuman AI
- Posts
- Robotics Special: A shared brain for self-driving cars
Robotics Special: A shared brain for self-driving cars

Self-driving cars are getting their own social network — NYU researchers have built a new system that lets autonomous vehicles talk to each other while they’re on the road. Also, humanoid robots are donning a lab coat and working hand-in-hand with human scientists.
P.S. The Robotics Special is designed to help you stay on the cutting edge of the latest breakthroughs and products in the industry. Our regular AI and Tech updates will resume as usual on Monday.
WHAT’S NEXT
The most important news and breakthroughs in robotics this week

Source: South China Morning Post
Swarm Squad: In a world-first, Chinese robotics firm UBTech’s Walker S1 humanoids just pulled off multi-humanoid coordination. Powered by the company’s BrainNet system, they used a "super brain" and "sub-brain" to dynamically share skills and optimize teamwork on production lines. This could potentially set the stage for fully autonomous factories that are faster, smarter, and more efficient. Watch them here.
Brain in a Box: Australian startup Cortical Labs just launched the world’s first commercial bio-computer, CL1, powered by living human brain cells. The $35,000 system uses lab-grown neurons wired onto a chip, allowing it to learn and adapt faster than traditional silicon-based AI, while consuming far less energy. The tech blurs the line between biology and computing, and could potentially revolutionize AI, robotics, and computing.
Factory Fresh: Boston Dynamics' Atlas humanoid is gearing up for real-world use cases, starting with part sequencing at Hyundai facilities. Using Reinforcement Learning and Foundation Models, Atlas can now pick, carry, and place objects with minimal human input — a major shift from flashy parkour stunts to practical industrial automation. The humanoid adapts on the fly, minimizing errors and reducing human strain.
Data Drive: NYU researchers just launched AnySense, a groundbreaking iOS app designed to gather visual training data for generalizable robotics models, making robot learning much more efficient. The app pairs with “The Stick,” a 3D-printed gripper, to capture high-quality training data. It even integrates with AnySkin, a touch-sensitive pad that provides feedback for better robotic grip — a potential game-changer for AI-powered robotics.
PRESENTED BY GUIDDE
Ideal for HR/Sales: Turn docs into step-by-step video tutorials instantly

Instead of sending complex PDFs and slide decks, turn them into professional How-To videos instantly with Guidde, the no-cost documentation tool.
It’s very simple and takes less than 5 minutes. Just upload a PDF, slide deck, or screen recording, and Guidde will create a video complete with captions, narration, and your logos.
Optimize your content with instant revisions
Share guides in any language

Source: IoT World
For years, autonomous vehicles (AVs) have had a huge communication problem. There was a massive bottleneck in their ability to share knowledge with each other, only exchanging information when they came into direct contact — like two strangers exchanging travel tips in passing. This meant that if an AV had only driven around in Manhattan, it wouldn’t have a clue about the road conditions in Brooklyn, and vice versa. Most AI learning systems also rely on centralized servers, causing major delays, inefficiencies, and data privacy concerns.
NYU Tandon’s research team decided to make these cars street smart. They developed Cached Decentralized Federated Learning (Cached-DFL), a method that lets AVs swap intel on road conditions through indirect model sharing, allowing AVs to distribute information about road conditions, signals, and traffic while maintaining data privacy. Even if two AVs never meet, their knowledge can still transmit to each other through a chain of vehicle relays, expanding the collective intelligence of autonomous fleets.
The system works like a digital gossip chain:
Cars within 100 meters share AI models via high-speed device-to-device communication
Each vehicle keeps a cache of up to 10 external models and refreshes its AI every 2 minutes
The system automatically purges outdated information based on a "staleness threshold"
Knowledge spreads far beyond immediate interactions, creating a network of shared experiences
Why this matters: The new method enables self-driving cars and other intelligent systems to learn collectively in real-time while preserving data privacy. Its decentralized approach boosts adaptability, scalability, and efficiency across AVs, drones, and robotic networks.
ROBOTS IN ACTION
How robots are transforming the world around us

Source: CNN
Digital Deputy: Humanoids are officially walking the beat in Shenzhen, China. EngineAI’s PM01 robot is assisting police patrols — a major step toward fully robotic policing. The robots come with agile movement, AI-powered perception, and open-source capabilities. Watch them in action here.
Plant Patrol: Gecko Robotics and NAES are teaming up in a $100M deal to modernize US power plants with AI-driven robots. The Cantilever platform will optimize plant performance, predict maintenance needs, and reduce costly downtime. With energy demand set to surge, automation could be key to keeping the grid stable.
Robotaxi Revolution: Uber riders in Austin can now get matched with Waymo's self-driving robotaxis as part of a new partnership. Pricing stays the same whether you ride with a human or AI, with plans in place to scale up in Austin and Atlanta soon.
Bay Bouncers: Robotics startup Splash is developing autonomous surface vessels to guard maritime borders and critical infrastructure from foreign threats. These AI-powered boats can identify, classify, and react to suspicious activity without human intervention, and have already logged 100+ miles of autonomous navigation in San Francisco Bay.
Senior Support: Tokyo researchers are rolling out AIREC, a humanoid robot designed to assist elderly patients in essential caregiving tasks. Using AI-powered motion control, AIREC adapts to patients' needs while reducing strain on human caregivers. If successful, AIREC could redefine elderly care by 2030.
PRESENTED BY YOU. COM
The next generation of deep research is here

You. com just launched ARI—the world’s first professional grade deep research agent. With a simple prompt, ARI will:
Analyze up to 400 sources simultaneously (10x+ more than others)
Produce polished PDFs in < 5mins (3x faster than others)
Generate rich visuals, charts, and clickable citations
PS: Superhuman readers can get 1 year free of you. com Pro ($180 value), which gives you priority access to ARI.
ROBO REEL
Watch: ‘Scientist’ humanoid makes itself useful in the lab
The next Nobel Prize in medicine could go to a humanoid.
Insilico Medicine, a U.S. clinical-stage biotechnology company, has introduced "Supervisor," a humanoid robot designed to assist in AI-driven, fully automated drug discovery. This marks the first instance of a bipedal humanoid robot being deployed in such a setting.
Why it matters:
Research cycles speed up dramatically and human error decreases significantly
24/7 operations become possible, boosting efficiency
Scientists can focus on creative problem-solving instead of tedious chores like pipetting
What’s trending in Robotics on socials this week
⚙️ Labor Leap: A Shanghai robot factory where humanoid robots are now in mass production just blew up on Reddit. The robots handle tasks in areas ranging from sales to heavy-load transport. Some users are excited, while others find it scary. Watch them in action here.
🚦 City Cruiser: Chinese journalist Li Zexin posted a video of a self-driving car, claiming the country now leads in tech innovation. The self-driving system already seems to be making city navigation easy, especially for rookie drivers facing tight streets and oncoming trucks.
🥋 Robot Rumble: If your house gets broken into, your Unitree humanoid robot won’t be dialing 911. The company posted footage of the robot carrying out a 720° roundhouse kick, and it’s got X users smashing the like button.
⚡ Betting Big: In a recent podcast, Figure AI CEO Brett Adcock made a bold prediction, claiming that a humanoid company will become the largest company in the world within our lifetime. The claim has divided tech socials.
🔧 Maker Magic: Amateur builder Lewis RK just posted his most recent invention, and it’s got X users buzzing. He 3D-printed a Fuchikoma Microbit caster wheel robot and now plans to include ultrasonic sensors for self-navigation. Watch it roll around here.
MARKET MOVEMENTS
Acquisitions, investments, funding, and more

Source: Shield AI
Here are the biggest developments in the robotics space that you should know about:
Defense startup Shield AI just raised $240M at a $5.3B valuation to commercialize its AI drone tech, making it the second-largest US defense tech startup after Anduril. Backers include L3Harris, Hanwha Aerospace, and Andreessen Horowitz.
Israeli startup CaPow has raised $15M to scale its Power-in-Motion system, which wirelessly powers robots as they move — getting rid of downtime. With Toyota Ventures leading the investment, CaPow is gearing up for global expansion.
Hyundai is partnering with Yandex spinoff Avride to develop autonomous robotaxis, set to hit Dallas later this year via Uber. The partnership deepens Hyundai’s push into self-driving technology.
Viam just secured $30M to scale its cloud-powered robotics development platform, making automation easier for companies without massive engineering teams. Plans are in place to expand in Europe.
Chinese robotics firm LimX Dynamics just secured $69M in Series A+ funding, with backing from Alibaba Group and other industrial players. This fresh capital will help scale production and accelerate AI-driven robotics innovation.
Your opinion matters!
You’re the reason our team spends hundreds of hours every week researching and writing this email. Please let us know what you thought of today’s email to help us create better emails for you.
What did you think of today's email?Your feedback helps me create better emails for you! |
Until next time,
Zain and the Superhuman AI team
FROM THE FRONTIER