Kiệt Đinh

@johncdwyckoff @PrismaXai tt manh len 😁

@kiet502 @PrismaXai gprisma kietlu

Over the past year, something fundamental has changed in robotics. And according to @PrismaXai 's latest industry memo, most people haven't fully realized it yet. For years, the robotics industry was obsessed with hardware. Who had the fastest humanoid. Who built the strongest robot arm. Who could create the most impressive demonstration. But after following the developments from IROS 2025, ICRA 2026, Stanford Robotics Center, and the broader Physical AI ecosystem, one thing has become increasingly clear: The future of robotics is no longer a hardware problem. It is a data problem. This may sound surprising at first. After all, robotics startups raised nearly $14 billion in 2025, more than the previous cycle peak. Humanoids are everywhere. Robot arms are cheaper than ever. Supply chains have matured. According to PrismaX, a complete robotics data collection cell can now be built for under $4,000 and deployed almost anywhere in the world. That statement alone tells us how dramatically the industry has evolved. Hardware is becoming accessible. Hardware is becoming standardized. Hardware is becoming a commodity. And when a technology becomes accessible to everyone, competitive advantage moves somewhere else. In the case of Physical AI, it moves to data. This is exactly why PrismaX's vision is so compelling. While many companies are racing to build robots, PrismaX is building the infrastructure that teaches robots how to learn. The company repeatedly emphasizes a simple idea: Better models start with better data. But in robotics, collecting data is far more complicated than collecting text or images. A robot does not learn from words alone. It learns from interaction. It learns from movement. It learns from force. It learns from success and failure. It learns from experience. This is why PrismaX continues to invest heavily in teleoperation. Every teleoperation session captures something incredibly valuable: Human judgment. Human intuition. Human problem-solving. Human adaptation. The countless micro-decisions people make naturally while interacting with the physical world. These experiences become action data. And action data becomes intelligence. What makes PrismaX particularly interesting is that the team does not view teleoperation and video data as competing approaches. Instead, they see the future in combining both. Video-only training has made a strong comeback across the industry. Human video provides scale. Teleoperation provides precision. Video provides breadth. Teleoperation provides quality. The future foundation models of robotics will likely require both. PrismaX describes this as a cross-embodiment future, where different forms of data contribute to a shared learning system. This perspective reflects a deeper understanding of where Physical AI is heading. Because the challenge is no longer teaching robots how to move. The challenge is teaching robots how to understand. Recent demonstrations at ICRA 2026 showcased robots pouring popcorn, folding shirts, stacking blocks, and performing increasingly complex manipulation tasks. The demonstrations were impressive. But PrismaX raised an important question: Can the robot do it again? And again? And again? Five hundred times? Without supervision? In a room that nobody prepared beforehand? This is the true test of intelligence. Not a perfect demonstration. But reliable performance in an imperfect world. And solving that problem requires something the robotics industry is only beginning to appreciate: High-quality data. Not all robotics data is valuable. Not all robotics data teaches the right lessons. Not all robotics data can produce useful foundation models. This is why PrismaX has repeatedly challenged one of the biggest trends in the industry: The sudden emergence of countless "robotics data companies." As Physical AI becomes the hottest sector in technology, many new entrants have rushed to sell data. Yet PrismaX argues that most of these companies lack deep robotics expertise. And without understanding robotics, it becomes difficult to understand what data actually matters. Because data quality is not measured by quantity. It is measured by usefulness. The difference is enormous. A robot performing a backflip requires one type of data. A robot completing real-world tasks in unpredictable environments requires something entirely different. This distinction sits at the center of @PrismaXai 's strategy. The company is not focused on creating viral robotics moments. It is focused on building the learning layer of Physical AI. The invisible infrastructure beneath the industry. The systems that collect experience. The systems that evaluate quality. The systems that transform human actions into machine intelligence. History shows that transformative industries are rarely defined by their most visible products. The internet was built on networking infrastructure. Cloud computing enabled modern software. Large language models were enabled by massive datasets. And Physical AI will be enabled by data infrastructure. That is why PrismaX's work feels increasingly important. They are not merely building tools for today's robots. They are building the foundation upon which future robotics intelligence may emerge. The next decade of robotics will not be won by the robot with the most impressive demo. It will be won by the companies that understand how intelligence is created. How experience becomes data. How data becomes learning. And how learning becomes autonomous capability. From teleoperation networks to data evaluation systems, from training frameworks to foundation model infrastructure, @PrismaXai is positioning itself at the very center of that transformation. The age of Physical AI is no longer coming. It has already begun. And the companies building the data layer today may ultimately shape the intelligence of every robot tomorrow 🦾 #PrismaX #PhysicalAI #Robotics #AI #Teleoperation #EmbodiedAI #FoundationModels @vivianrobotics @johncdwyckoff

@kiet502 Vẽ tay đẹp quá bro

The most important message from ICRA 2026 was not that robots can pour popcorn. It was not that robots can fold shirts. It was not that robots can stack blocks. The real message was something much deeper. Robots are finally becoming capable enough that hardware is no longer the main bottleneck. And that realization sits at the heart of everything PrismaX is building. For years, the robotics industry competed on demonstrations. Who could make a robot walk better. Who could make a robot move faster. Who could create the most impressive viral video. Today, many companies can do that. The challenge has changed. Now the question is: Can the robot repeat the task 500 times? Can it adapt when the environment changes? Can it succeed when nobody carefully prepares the room beforehand? Can it learn from experience? This is where @PrismaXai 's vision becomes incredibly important. The future of Physical AI will not be determined by hardware alone. It will be determined by data. Because every robot that succeeds in the real world is standing on top of thousands of previous experiences. Every grasp. Every movement. Every correction. Every mistake. Every successful task. These interactions create the data that teaches robots how to understand reality. And PrismaX is building the infrastructure that makes this possible. Through teleoperation, PrismaX allows human operators to generate high-quality robotics data in real-world environments. Through data collection systems, those interactions become structured learning signals. Through data scoring and evaluation, human judgment helps identify what good performance actually looks like. Through training frameworks, those experiences can be transformed into the next generation of robotics foundation models. This is why I believe @PrismaXai is positioning itself at one of the most important layers of the Physical AI stack. Not at the hardware layer. Not at the demo layer. But at the learning layer. The layer where intelligence is created. Recently, PrismaX has repeatedly emphasized a simple but powerful idea: Better models start with better data. That statement may ultimately define the future of robotics. Because the industry is beginning to realize that intelligence does not emerge from robots alone. It emerges from experience. And experience must be collected, evaluated, organized, and transformed into knowledge. The robots showcased at ICRA 2026 demonstrate what machines can do today. PrismaX is focused on a much bigger question: How do we build systems that allow robots to become smarter tomorrow? That is why the company's investments in teleoperation, robotics data infrastructure, human scoring, and foundation model training are so significant. They are not optimizing for the next demonstration. They are building for the next decade. A decade where Physical AI moves beyond controlled environments and enters factories, warehouses, businesses, and eventually everyday life. When that future arrives, the most valuable asset may not be the robot itself. It may be the data and learning infrastructure that taught the robot how to operate in the first place. And from everything @PrismaXai has shared so far, it is clear that the team understands this transformation better than most. The future of robotics will not belong to the robot that performs perfectly once on stage. It will belong to the systems that enable robots to learn continuously, adapt endlessly, and succeed thousands of times in the real world. That is the future PrismaX is helping build 🦾 #PrismaX #PhysicalAI #Robotics #AI #Teleoperation #EmbodiedAI #ICRA2026

@Hanviiw_JS Từ chối chia sẻ kk

@kiet502 Đã tiếp thu

@eryxne03 Tks sir

@kiet502 Vẽ tay vậy đẹp quá bro ơi

@DngDZ16 @PrismaXai Lên 7 luôn nhé

Happy 1st Birthday, @PrismaXai 🎉 One year may not seem like a very long time, but my journey with @PrismaXai has been filled with unforgettable memories, valuable lessons, and amazing people. I am currently a Stabilized member, and when I first joined @PrismaXai, my goal was simple: to learn more about Web3 and become part of an active and passionate community. However, as time went by, @PrismaXai gave me much more than I ever expected. One of the most memorable moments in my journey was seeing some of the content I worked hard to create receive recognition from @vivianrobotics. Several of my posts were liked by her, and one of them was even quoted and shared. To some people, this might seem like a small interaction, but for me, it meant a lot. It was a reminder that dedication and effort never go unnoticed. That recognition gave me confidence, motivation, and a stronger desire to continue contributing to the community. I would also like to express my sincere gratitude to Ambassador @johncdwyckoff . From my very first days in @PrismaXai, when I was still learning and trying to find my place, he was always there to encourage, guide, and support me. His advice and kindness helped me grow more confident and become more involved within the community. I truly appreciate everything he has done for me. What I love most about @PrismaXai is not only the opportunities to learn and contribute, but also the incredible community culture that has been built here. The feedback review sessions are always thoughtful and detailed. Every submission receives genuine attention, helping members improve their skills and become better content creators. Those sessions have taught me a lot and played an important role in my personal growth. Another highlight for me has been the weekly karaoke gatherings. They may seem like simple community activities, but they bring people together in a special way. They create opportunities for members to connect beyond work and tasks, share laughs, build friendships, and strengthen the bonds that make this community feel like home. Looking back on this past year, I realize that the greatest value @PrismaXai has given me is not just knowledge, rewards, or achievements. It is the friendships, the support, the encouragement, and the feeling of belonging to a community that genuinely cares about its members and their growth. Thank you, @PrismaXai, for an incredible first year. Thank you to everyone who has been part of this journey, whether through collaboration, support, friendship, or inspiration. I am grateful for every experience and every memory we have created together. This is only the beginning, and I truly believe even greater things are ahead for @PrismaXai and its community. Wishing @PrismaXai continued success, growth, and many more milestones in the years to come. 💜🚀 Happy 1st Birthday, @PrismaXai 😘 @vivianrobotics @MaxC16134 #HappyBirthdayPrismaX #MyStoryWithPrismaX

@kun090120 @PrismaXai @johncdwyckoff @MaxC16134 @vivianrobotics

Happy first birthday, @PrismaXai! 🍻🎂💔 May PrismaX continue to grow and thrive on its chosen path. I and the PrismaX community will always be with you. @johncdwyckoff @MaxC16134 @vivianrobotics Happy Birthday! 🥳🎂 #HappyBirthdayPrismax #Prismax #Robotics

@eryxne03 chia sẻ tâm tình hay quá

Happy Birthday, PrismaX !🎆 One year ago, PrismaX was still an idea that many people found difficult to fully understand. The concept of connecting humans, robots, and AI through real world data felt ambitious, and for many, the vision of Physical AI still seemed far away. Fast forward to today, and that vision feels more tangible than ever. PrismaX has grown into a community of builders, contributors, and believers who understand that the future of AI will not be limited to conversations on a screen. The next generation of intelligence will need to interact with the physical world, learn from real environments, and continuously improve through real world experience. What I appreciate most about PrismaX is that it is tackling one of the biggest challenges in robotics. Data. While language models can learn from vast amounts of information online, robots need something completely different. They need human actions, real tasks, and real world interactions. PrismaX is creating a system where anyone can contribute to that process while helping build the foundation of Physical AI. Over the past year, the project has become more than just a platform. It has become a place where people learn about robotics, AI infrastructure, teleoperation, and the future of intelligent machines. Through community discussions, content creation, events, and collaboration, many contributors have played a role in turning an ambitious idea into a growing ecosystem. Every contribution matters. Every teleoperation session generates valuable data. Every piece of feedback helps improve the network. Every community member adds something unique to the journey. What makes PrismaX exciting is not only what has already been built, but what it has the potential to become in the years ahead. The first year was about proving the vision. The next years will be about scaling it. As Physical AI continues to move from concept to reality, projects focused on real world intelligence and robotic learning will become increasingly important, and PrismaX is positioning itself at the center of that transformation. Congratulations to the entire PrismaX team, every contributor, and every community member who has helped shape this journey so far. One year down. The real mission is only getting started. #HappyBirthdayPrismaX @PrismaXai | @MaxC16134 | @vivianrobotics @illuviumio_eth | @johncdwyckoff

@allscaleio @circle Wow, that's really great

Excited to share that we have joined the @Circle Alliance Program, a global community of teams focused on bringing the world on-chain, powered by USDC!

@ThuyTrang108 happy 1st birthday gPrisma

📖 My Story with PrismaX 📖 Happy 1st Birthday, PrismaX! A Humble Beginning, An Unbelievable Speed When I first joined PrismaX, I was just an unknown name among thousands of members, an anonymous Reactive starting from scratch: I didn't know anyone, didn't fully understand the ecosystem, but had enough curiosity to stay and learn. What truly interested me and made me want to be a long-term partner with PrismaX wasn't just the AI ​​technology or data. It was the vision of connecting robotics, open ecosystems, and people. PrismaX is aiming for a future where robots don't simply replace labor, but become tools to help humans better develop their creativity, adaptability, and solve complex problems. I not only want to follow that journey, but also want to be a part of contributing to the development of the PrismaX community and ecosystem in the future. The Journey Through the Progression Levels PrismaX's Progression system has 7 levels: Reactive → Assistive → Proactive → Exploratory → Stabilized → Navigational → Groundbreaker. Looking at the estimated numbers, the ratio of Groundbreaker to Reactive, I understand that each step forward is not a simple matter. Yet I've gone from 0 to Navigational ---> Role 6, just one step away from Groundbreaker. Each time I level up, I look back and ask myself: what have I done to deserve this? The answer is always: consistently showing up every day, contributing regularly, and constantly finding ways to help the community a little better. The two numbers I'm most proud of are: the number of active messages in Discord, not spam, but discussions, support for new members, and project information exchanges every day and 127 contributions recorded in the content-curators channel under the tag "JohnNguyen Contribute." 127 threads on X: telling the PrismaX story in my own way Most of my contributions revolve around writing threads and posts on X. I see each thread not just as a post to increase engagement, but as a way to "translate" updates, vision, and values ​​of PrismaX into a more accessible language for the community, especially the Vietnamese user community. Some posts I write to explain the project's mechanics to newcomers, some share personal experiences participating in events, and some threads simply spread positive energy whenever PrismaX reaches a new milestone. 127 listening sessions might seem like a number, but for me, it represents 127 times I chose to dedicate my time to contributing a small part to the shared story. Teleoperator: When I don't just type, but truly "touch" the technology. If I had to choose one role that made me feel truly part of PrismaX's mission, it would definitely be Teleoperator. This isn't a symbolic title; it's directly linked to the operation and data collection process for robots, a core part of the physical AI mission that PrismaX is building. The experience of being a Teleoperator gave me a completely different perspective than just reading about the project: I understand better why each operation, each session, is important for AI training, and why PrismaX isn't simply a crypto or social community, but a real project building a data foundation for robots in the real world. This role made the threads I wrote later more authentic, because I wrote from real experience, not just from information I read. A new journey is just the beginning. Looking back on this short but remarkable journey, from an unknown nobody to Navigational, from a silent member to someone who has written over a hundred shared posts, I realize that what PrismaX offers is more than just a role or a badge. It's the feeling of belonging to something that is truly being built. PrismaX stands out to me because of its vision that goes far beyond mere technology. The combination of robotics, AI, and an ecosystem built around people opens up opportunities to create practical applications with a significant impact. I not only want to follow that journey, but also want to be a part of contributing to the development of the PrismaX community and ecosystem in the future. #HappyBirthdayPrismaX #MyStoryWithPrismaX @PrismaXai @vivianrobotics @MaxC16134 @johncdwyckoff

@Hanviiw_JS chỉnh chu quá

🎂 Happy 1st Birthday, PrismaX! 🎂 A year ago, PrismaX was just an idea that many people were still trying to understand. Today, it has become a growing community united by a shared vision of Physical AI, robotics, and the belief that human intelligence still plays an essential role in shaping the future. Looking back, one of the things I appreciate most is not only the technology being built, but also the people I met along the way. Through discussions, content creation, community events, feedback, and countless conversations, PrismaX has given me the opportunity to learn more about robotics, AI infrastructure, and what it truly means to contribute to something bigger than myself. Every post. Every interaction. Every teleoperation session. Every contributor. They are all small pieces of a much larger story. As PrismaX celebrates its first birthday, I believe this is only the beginning. The vision is becoming clearer. The community is becoming stronger. And the future of Physical AI feels more real than ever. Thank you to the team, the contributors, and everyone who has been part of this journey. Happy Birthday, PrismaX. 🤎 Here's to many more years of building the future together. #HappyBirthdayPrismaX #MyStoryWithPrismaX @PrismaXai @MaxC16134 @vivianrobotics @illuviumio_eth @johncdwyckoff

@johncdwyckoff @PrismaXai chi tiết quá

GPrisma everyone 👋 I finally got hands-on with VLA Foundry (Beta) from @PrismaXai and after running multiple validations myself, I think most people are completely misunderstanding what this actually is. I attached a clip from one of the tasks I reviewed: Bimanual robot arms arranging flowers into a vase — simple on the surface, but surprisingly rich in training signal. At first glance, VLA Foundry looks like a typical “watch & rate” platform. But once you actually do it, you realize: 👉 This is NOT about watching robots. 👉 This is about defining what “good data” means for physical AI. Honestly, I didn’t expect this when I first tried it. As validators, we don’t just score randomly. We’re effectively acting as a human judgment layer for robot learning. Each task is evaluated based on: • Camera coverage & clarity (especially overhead + wrist views) • Whether the task truly matches the prompt • Execution speed (too slow = bad signal) • Precision & control • Smoothness of motion (this matters more than you think) But here’s what changed my perspective: Most robotics datasets today are messy, inconsistent, and honestly… not very useful for training real-world AI. What @PrismaXai is doing differently: 👉 Setting a strict standard for what counts as usable robotics data 👉 Filtering out anything that doesn’t meet that bar 👉 Reinforcing it with human validation + consensus scoring From my own testing so far: • Clean runs with clear intent = easy to score • Slight hesitation or micro-adjustments = immediately noticeable • Overlapping actions (both arms doing too much at once) = reduces quality You start to see the difference between: 👉 data that trains a model vs data that just fills storage This part actually surprised me the most. Another interesting mechanic: Each dataset is reviewed by multiple validators, and your rewards depend on how close your evaluation is to the consensus. So it’s not just about “your opinion” — it’s about aligning with a shared understanding of quality. After numerous tests, one thing is clear: 👉 This is not just a feature. 👉 This is the foundation for how @PrismaXai scales robot training data. And if they execute this well, this could become a really important layer in the Physical AI stack. Public Beta is coming soon. If you're exploring this early, I highly recommend joining the community — there are already valuable discussions around validation strategies and scoring patterns: discord.gg/aRFTMXSu You're not just testing a product — you're helping shape how robots learn. Do you trust human judgment to shape the behavior of AI?

@Kyo_NLD @PrismaXai @a16z Chúc mừng PrismaX

Hãy cùng nhìn lại 1 năm của Robotics và PrismaX😍 Thời gian trôi qua rất nhanh , vậy là nửa năm 2026 đã đi qua và hành trình phát triển của @PrismaXai vẫn tiếp tục Rất nhiều sự kiện đáng chú ý , đặc biệt là tuần lễ công nghệ được tổ chức tại NY do @a16z điều hành , với sự góp mặt của các tập đoàn công nghệ cũng như các dự án Robotics như : @openmind_agi , đã đem đến những sản phẩm về Physical thật ấn tượng PrismaX cũng đã góp mặt trong sự kiện lần này , và cũng thảo luận về AI + Data + Robotics 🤖 Sự trưởng thành về phần cứng đã được khắc phục - có 1 chuỗi cung ứng các thiết bị máy ổn định hơn Từ việc lắng nghe những khách hàng của mình họ đã thay đổi mở rộng dữ liệu teleoperation thay vì chỉ đào tạo robot qua video như trước - bởi vì teleoperation hoạt động rất tốt thu thập được chất lượng hoạt động đạt kết quả cao Trong tương lai , tác giả cho rằng ngành sẽ tiến tới các pipeline cross-embodiment kết hợp cả hai Bên cạnh đó , làn sóng về physical AI cũng đã xuất hiện rất nhiều các dự án cạnh tranh - Dữ liệu trở thành ngành kinh doanh sôi động PrismaX đã công bố tiêu chuẩn đánh giá dữ liệu của riêng họ (tốc độ, độ trơn mượt, đa dạng, khả năng xử lý từng việc một cách gọn gàng) Kì vọng sắp tới , sẽ sớm đưa ra được những giải pháp về doanh thu khép kín thực sự từ thị trường và Robot chính thức bước vào sự cạnh tranh khốc liệt . @vivianrobotics @castorhat @MaxC16134 🫡🇻🇳

@hoaihoai69232 @PrismaXai đầy đủ chi tiết

@PrismaXai Some technologies are created to make life easier. Others are created to redefine the relationship between humans and the world around them. Robotics and Physical AI belong to the second category. Looking at the innovations showcased at ICRA 2026, what stands out is not simply smarter robots, more capable hardware, or increasingly powerful AI models. What truly inspires is the realization that the boundary between artificial intelligence and the physical world is slowly disappearing. For years, AI lived mostly behind screens, processing text, images, and digital information. But the future being built today is very different. It is a future where intelligent systems can observe their surroundings, understand context, make decisions, and act within the real world. Every robot that learns to balance. Every robotic arm that learns to grasp an object. Every system that learns to collaborate with humans. Represents a small step forward with enormous implications. Behind every movement lies countless hours of research, experimentation, and real world data. Perhaps the most exciting part is that we are no longer merely imagining the future. We are actively building it. Events like ICRA 2026 remind us that Physical AI is no longer science fiction. It is becoming reality, one breakthrough at a time. And years from now, we may look back on this moment and realize that we were witnessing the opening chapters of a new technological revolution. A revolution where AI does not simply think. It sees. It learns. And it acts alongside humanity. 🤖✨ #PrismaX

@Hanviiw_JS khỏi mất lòng rồi

@kiet502 khóc đi sửa hình đăng lại cho kk

When people look at this logo, they see PrismaX. When I look at it, I see something else. Hundreds of contributors. Thousands of interactions. Countless hours spent helping Physical AI move forward. Every operator. Every creator. Every community member. A small piece on their own. But together, they become something much bigger. Hundreds of voices. One mission. 🤎 @PrismaXai #PrismaXai #PhysicalAI @vivianrobotics @MaxC16134

@BonBon380991 xịn quá

Amidst busy work hours, sometimes all we need is a moment of quiet to recharge. A few minutes of rest and exploring the interesting things from the Dlicom community. Technology is not just about speed of development or growth figures. It's also about connecting, learning, and sharing values ​​with like-minded people. @DlicomApp @dlicom_vn

@tunct101 @PrismaXai Kinh nghiệm quá

Hello everyone, another control session is starting, here are a few things I want to share about improving Teleoperation on @PrismaXai: > tip1: First, don't look at the timer too much; it only makes you nervous and more prone to errors. > tip2: Staying calm is most important; rushing only leads to mistakes. > tip3: Utilize the shape of the tool instead of trying to force it. > tip4: Points like holes, handles, and edges are always much easier to grab than the main surface, for example, the tail of a teddy bear or the handle of a cup. > tip5: If the object is far away, don't try to reach it; use another tool to push or pull it closer first. > tip6: Teleoperation is actually real-time situational handling; there's no fixed method. >> The important thing is that both failures and successes are useful because the robot learns from all human behavior, not just from correct actions. So what are you waiting for? Try this teleoperation experience now! app.prismax.ai/robots-center @PrismaXai @MaxC16134 @vivianrobotics