Gemini Robotics 1.5, launched by Google DeepMind, marks the beginning of agentic robotics. Robots that can: Search local rules before sorting your trash. Fold your laundry by color. Plan, adapt, and explain why they act - all in plain language. This is Gemini’s brain inside a body. A fusion of vision, language, and motion, so robots can reason, plan, and act across any embodiment. What makes it wild: -Two brains in sync - one for reasoning, one for movement. -Cross-embodiment learning - teach one robot, all others learn. -Tool use - can query Google Search mid-task. Safety by design - passes Google’s ASIMOV benchmark for alignment. Just think about what that means…

Google DeepMind's Gemini Robotics 1.5 heralds a new era in agentic robotics, introducing machines with the capacity to think before they act. This groundbreaking development moves beyond simple task execution, enabling robots to reason, plan, and adapt in the physical world. Imagine robots capable of understanding local waste disposal regulations before sorting recyclables, or folding laundry with an awareness of color coordination. This fusion of vision, language, and motion allows robots to process information and articulate their decision-making in plain language, effectively bridging the gap between digital intelligence and physical interaction. The Gemini Robotics architecture is particularly notable for its dual-processing system, synchronizing a reasoning engine with a movement controller. Furthermore, its cross-embodiment learning capability means that knowledge gained by one robot can be shared across the entire network, drastically accelerating development and deployment. The ability for robots to utilize tools, such as querying Google Search mid-task, and their inherent safety design, validated by Google's ASIMOV benchmark, underscore the sophisticated nature of this advancement. This strategic move positions Google as a significant contender in the embodied AI space, directly engaging with other leading efforts like Tesla's Optimus and Figure. The implications are vast, suggesting a future where robots are not just automated tools, but intelligent agents capable of handling complex, unexpected situations with logical explanations. This is more than just an incremental upgrade; it represents a fundamental shift in how we envision robotics and artificial intelligence interacting with our environment. Robotics ArtificialIntelligence EmbodiedAI

Stunning! But in which framework of behaviouristic and consequentional rules does the robot think? We need to urge the implementation of universal a ruleset which applies to every AI if it has the ability to think and act autonomously.

In case you missed it, Google recently launched Gemini Robotics, giving robots the ability to think before they move 😳 Gemini Robotics 1.5, launched by Google DeepMind, marks the beginning of agentic robotics 🤖 Robots that can: → Search local rules before sorting your trash → Fold your laundry by color → Plan, adapt, and explain why they act - all in plain language. This is Gemini’s brain inside a body. A fusion of vision, language, and motion, so robots can reason, plan, and act across any embodiment. What makes it wild: ↳ Two brains in sync - one for reasoning, one for movement. ↳ Cross-embodiment learning - teach one robot, all others learn. ↳ Tool use - can query Google Search mid-task. ↳ Safety by design - passes Google’s ASIMOV benchmark for alignment. Just think about what that means: Robots that understand context, handle the unexpected, and talk through their logic. That’s basically the bridge between AI chatbots and the physical world. Google is now quietly positioning itself as the leader of embodied AI, and thus going head-to-head with Tesla’s Optimus or Figure. The AI arms race is getting physical. Thanks Linas Beliūnas

🤖 Can robots truly learn like humans — not just perform tasks? Something big is happening in the world of robotics & AI. 🚀 Amazon’s latest Frontier AI research might be a major turning point, taking us closer to that reality than ever before. 🔹 Learning Beyond Code: Instead of being programmed for specific actions, these robots can learn by doing — observing, adapting, and improving through trial and feedback. 🔹 Generative + Physical Intelligence: A new wave of AI that merges generative reasoning with robotic embodiment — allowing systems to not just predict text, but also move, sense, and react in the real world. 🔹 Continuous Adaptation: Robots can now update their own behavior in real time, refining their understanding of space, force, and coordination. 🔹 Multi-Agent Collaboration: Multiple robots working together — sharing data, adapting to each other’s actions, and achieving shared goals with minimal human intervention. 🔹 Real-World Deployment: These advances aren’t staying in labs — they’re being tested across logistics, retail, and warehouse automation where learning agility = efficiency. ⚙ Why This Matters The line between automation and autonomy is fading. We’re moving from “robots that follow instructions” → to “robots that think, predict, and decide on their own.” That shift could redefine: ✔ How factories run ✔ How humans and robots collaborate ✔ What ‘work’ even means in the age of intelligent machines 🧠 The Big Question If robots begin to self-learn and adapt, who sets the boundaries? Ethics, safety, and reliability will now depend as much on data integrity as on design. Are we ready for self-learning, autonomous robots working alongside humans? Drop your perspective below 👇 — breakthrough or risk? Let’s discuss. 👉 Follow 1stepGrow Academy for more on AI, robotics, and the career skills that matter in the age of embodied intelligence. #AmazonAI #FrontierAI #Robotics #TechInnovation #AIresearch #1stepGrow

🤯 Google nailed it. Gemini Robotics 1.5 just gave machines a mind of their own. Google DeepMind just dropped Gemini Robotics 1.5, and it’s not just another AI update it’s a shift in what it means for machines to act intelligently. For the first time, robots aren’t just following instructions they’re reasoning, planning, and even explaining why they act. Here’s how it works 👇 → Gemini Robotics 1.5 handles perception and movement - seeing, planning, and executing complex tasks. → Gemini Robotics-ER 1.5 acts as the robot’s “thinking brain,” making plans, calling tools like Google Search, and coordinating multi-step missions. Together, they enable robots that can: → Check local rules before acting. → Fold laundry by color. → Plan, adapt, and explain their choices - in plain language. This is the dawn of agentic robotics; where vision, language, and motion fuse into machines that can learn, reason, and act across any form. Why it’s a big deal: → Cross-learning: Teach one robot, and others learn instantly. → Safety by design: Built-in safeguards and alignment with Google’s AI safety standards. →Tool use: Robots can query Google Search mid-task. This is more than automation; it’s the birth of autonomous reasoning. Are we ready to collaborate with intelligence that can think for itself?

🤖 AI vs Robots: What Most People Still Get Wrong Let’s settle this once and for all because I hear this everywhere: “AI and robots are the same thing.” Nope. Not even close. Here’s the simplest way to think about it 👇 AI is the brain 🧠 Robotics is the body 🤖 AI (Artificial Intelligence) is the thinking, the part that analyzes, learns, decides, and adapts. It’s the software, the logic, the reasoning engine. Robotics, on the other hand, is the hardware the physical body that moves, picks things up, or follows instructions. Now, here’s where people mix them up: When you see a robot acting “smart,” that’s because AI has been placed inside the robot to guide its actions. But a robot, by itself, isn’t “intelligent.” It’s just following a script, like a puppet waiting for a puppeteer. So when people say, “AI and robots will take over the world,” I smile a little. Because we’re not even close to that yet. We’re still building the bridges between those two worlds ; AI (the brain) and robotics (the body). And here’s the truth: You don’t need to fear it. You just need to understand it. Because once you grasp how these systems work, you’ll see opportunities, not threats. Opportunities to make your work faster. Your life simpler. And your kid’s futures, smarter. AI isn’t about coding or becoming techy overnight. It’s about learning to use the tools that make you more useful, relevant, and unstoppable. So the next time someone says, “Robots are coming for us,” just smile and say, “Not without a brain, they aren’t.” 😉 ✨ I’m Marina , I simplify AI so you can bulletproof your mind, career, and creativity. ❓❓ Be honest, before reading this, did you think AI and robotics were the same thing?😅 . . … #AIExplained #AIForBeginners #DigitalLearning #FutureOfWork #AIInSimpleTerms #AIForParents #AIEducation #MidlifeReinvention #TechMadeSimple

🤖 𝐓𝐨𝐩 5 𝐇𝐮𝐦𝐚𝐧𝐨𝐢𝐝 𝐑𝐨𝐛𝐨𝐭𝐬 𝐨𝐟 2025 🚀 — 𝐓𝐡𝐞 𝐃𝐚𝐰𝐧 𝐨𝐟 𝐭𝐡𝐞 𝐇𝐮𝐦𝐚𝐧-𝐀𝐈 𝐖𝐨𝐫𝐤𝐟𝐨𝐫𝐜𝐞 𝐈𝐬 𝐇𝐞𝐫𝐞! 🦾 The line between man and machine is blurring faster than ever. From factory floors to research labs, humanoid robots are no longer science fiction — they’re joining the workforce. Here are the 5 humanoids leading the next industrial revolution 👇 🦾 1. 𝐎𝐩𝐭𝐢𝐦𝐮𝐬 (𝐓𝐞𝐬𝐥𝐚 𝐁𝐨𝐭) Built by Tesla under Elon Musk, Optimus is designed to handle repetitive and risky tasks. It can lift over 20 kg, assist in factories and homes, and learn new skills via AI. Expected to enter internal use in 2025 and mass rollout by 2027. 👁️ 2. 𝐀𝐦𝐞𝐜𝐚 (𝐄𝐧𝐠𝐢𝐧𝐞𝐞𝐫𝐞𝐝 𝐀𝐫𝐭𝐬) The most human-like face in robotics. Capable of emotional reactions, conversations, and powered by GPT-style AI models, Ameca leads in AI-human communication research. ⚙️ 3. 𝐀𝐭𝐥𝐚𝐬 (𝐁𝐨𝐬𝐭𝐨𝐧 𝐃𝐲𝐧𝐚𝐦𝐢𝐜𝐬) Fully electric and agile — Atlas can jump, run, and manipulate heavy objects. It’s being tested for industrial assistance and disaster response scenarios. Boston Dynamics calls it their “most dynamic humanoid ever.” 📦 4. 𝐃𝐢𝐠𝐢𝐭 (𝐀𝐠𝐢𝐥𝐢𝐭𝐲 𝐑𝐨𝐛𝐨𝐭𝐢𝐜𝐬) Your warehouse assistant of the future. Digit is built to lift, stack, and walk among humans, already working with Amazon and Ford to optimize logistics. Commercial rollout scaling through 2026. 🧠 5. 𝐆1 (𝐔𝐧𝐢𝐭𝐫𝐞𝐞 𝐑𝐨𝐛𝐨𝐭𝐢𝐜𝐬) A humanoid that costs less than a used car (~₹13 lakh / $16,000). Compact, research-ready, and AI-powered — it’s making robotics accessible to everyone. Adopted by global universities and labs for R&D. Used in museums, labs, and tech expos globally. 🧩 𝐖𝐡𝐲 𝐢𝐭 𝐦𝐚𝐭𝐭𝐞𝐫𝐬: Humanoids are moving from demonstrations to deployments. By 2030, analysts expect over 1 million humanoid robots in daily industrial use. This is not about replacing humans — it’s about redefining work. 👇 𝐃𝐫𝐨𝐩 𝐲𝐨𝐮𝐫 𝐩𝐢𝐜𝐤 𝐢𝐧 𝐭𝐡𝐞 𝐜𝐨𝐦𝐦𝐞𝐧𝐭𝐬 — 𝐥𝐞𝐭’𝐬 𝐬𝐞𝐞 𝐰𝐡𝐞𝐫𝐞 𝐭𝐡𝐞 𝐟𝐮𝐭𝐮𝐫𝐞 𝐬𝐭𝐚𝐧𝐝𝐬! 🔁 Save this post to track how fast humanoids are evolving. 📲 𝐅𝐨𝐥𝐥𝐨𝐰 @𝐍𝐞𝐨𝐒𝐡𝐚𝐫𝐗 𝐟𝐨𝐫 𝐝𝐚𝐢𝐥𝐲 𝐝𝐞𝐞𝐩 𝐝𝐢𝐯𝐞𝐬 𝐨𝐧 𝐀𝐈, 𝐫𝐨𝐛𝐨𝐭𝐢𝐜𝐬, 𝐚𝐧𝐝 𝐟𝐮𝐭𝐮𝐫𝐞 𝐭𝐞𝐜𝐡. #HumanoidRobots #AIRevolution #FutureOfWork #ArtificialIntelligence #RoboticsInnovation #Automation #TechTrends2025 #AIFuture #HumanRobotInteraction #DeepTech #Innovation #NextGenTech #IndianTech #StartupIndia #DigitalTransformation #NeoSharX #FutureIsNow #TechEvolution #SmartAutomation

Figure AI has unveiled Figure 03, its most advanced humanoid robot yet. Built around Helix, the company’s own vision-language-action (VLA) model, Figure 03 represents a complete redesign for real-world life, smarter AI, softer materials, and a system engineered for mass manufacturing. Figure’s mission has always been to build humanoids that can learn from people and adapt to the complexity of the world around them. With Figure 03, that vision finally steps off the lab floor and into daily life. At the core of this breakthrough is Helix, a new layer of sensory and motor intelligence designed for reasoning and real-world interaction. The new model was developed with four clear goals: deeper reasoning through Helix, safer design for home environments, scalable manufacturing, and readiness for global deployment. Figure 03 features a major upgrade in perception and dexterity. Its vision system offers twice the frame rate, one-quarter the latency, and a 60 percent wider field of view compared to previous models. Each fingertip can detect just three grams of pressure, about the weight of a paperclip, enabling highly precise “pixels-to-action” learning. To coexist safely and comfortably with people, Figure 03 has been designed with soft textiles and foam at pinch points, is nine percent lighter and smaller than Figure 02, and includes safer battery systems, wireless charging, and a more natural voice interface with improved audio. What truly sets Figure 03 apart is its scalability. Unlike most humanoids, it was engineered for mass production from the start, using simplified components, a die-cast structure, and an in-house supply chain. Figure’s new factory, BotQ, can produce 12,000 robots per year, with capacity to scale to 100,000 units. The prototype era, it seems, is over. With twice the actuator speed and smarter perception, Figure 03 can already take on tasks across logistics, retail, manufacturing, and healthcare. Its wireless charging and fleet data synchronization make continuous 24/7 humanoid operations possible, even down to customizable uniforms for different work environments. Figure 03 marks a defining moment for robotics. It’s a humanoid built not just to demonstrate technology, but to live and work alongside people. From the lab, to the home, to the world at scale, the future is finally standing on two feet.

Rodney Brooks' latest blog post–"Why Today's Humanoid Robots Won't Learn Dexterity”– highlights the major shortcomings in current approaches to solving robot dexterity, and validates many of Sanctuary AI’s core principles in our approach. “We have to get the humanoid robots to be able to do human-like manipulation as that is the key to them making both economic and technological sense … It is precisely the humanoids’ raison d’être.” – Rodney Brooks From the outset, we’ve believed that hands represent the highest-value component of the humanoid tech stack, and we’ve invested heavily in what we see as the true unlock - not just for humanoid utility and promise - but for any AI-enabled robot. Achieving human-level dexterity isn’t just a hardware, software, AI, sensor, controls or data challenge — it’s a systems-level problem that demands deep expertise across all domains, rarely found in Embodied AI startups. Robust, reliable, and high-performance hardware is just your entry ticket, then you need software infrastructure capable of ingesting and learning from sensor rich task data, and then tuning know-how to produce truly performant AI policies that customers care about. It is truly one of the most challenging technological problems of our generation. We are increasingly confident that vision data alone, paired with high DOF hands, is insufficient to surpass low-cost conventional gripper capabilities on most manipulation tasks. Tactile sensing is a 'must' for achieving genuine human-like dexterity. At Sanctuary AI, we are working with a wide range of end effectors to ensure customers get their problems solved today, and in the future. If you’re passionate about shaping the future of dexterous intelligence with physical AI, view our positions:https://lnkd.in/ggwBUQBH Learn more about our technology: https://lnkd.in/g8UvAU6D Link to Rodney Brooks’ blog posts: https://lnkd.in/gszxcWBi