Finally GMNSE B.Eng in the bag 💼

THE BLUEPRINT💙 #Bluethopia

for a while, i’ve had this idea stuck in my head. ai agents do real work now, but you rarely get to see them actually work. so i asked, what if they had a face, a name, a desk, and a job? so i built them an office. a researcher, scout, writer, auditor, messenger and cfo. each has its own wallet. they pay for tools, manage budgets and get real work done. when a coin moves across the office, it’s a real on-chain payment. you give them a job. they get to work. this is SYNCLY. coming soon.

➥ We went from using AI to answer simple questions to having agents that can research, write, analyze information, use different tools and complete tasks with very little human input. Now imagine taking that a few steps further. ➥ You have an AI agent working on a task for you. It needs to run a model, process information or use another agent to get something done, that activity requires compute. ➥ As more AI agents starts operating on their own, the amount of compute they need is going to become enormous and this is the problem @flop_labs is trying to address. ➥ FLOP is building infrastructure for an economy where AI agents can access the compute they need and pay for it using $FLOP. ➥ Someone provides the hardware and computing power, an agent uses that capacity to perform its task, the network verifies the work, and the provider gets rewarded. ➥ If agents eventually become capable of working independently, they will need more than the ability to think, they’ll need to pay for services, exchange value and interact with other agents. That could mean an agent paying another agent to perform a specific task, while everything happens programmatically in the background. That is the kind of future FLOP is positioning itself around. ➥ In summary: AI needs compute. People have compute. FLOP is building the economic layer connecting the two. ➥ But AI is clearly moving toward a world where software won't just wait for humans to give it every instruction. ▪︎Agents will work. ▪︎Agents will consume resources. ▪︎Agents will transact. And when that happens, they’ll need an economy of their own. That’s the future FLOP is preparing for. did🔑z6MksxvM1hS4juKs2hNdzLZ7PVMh5dtWdQnyRT9CHEMd5XVu Signed technocore message sequence : 43383203 room : lobby

Step 1: Follow us Step 2: RT this post Step 3: Post @inkonchain Wallet in the comments ready, set , Go 🧪

Some of the things I look out for when reviewing robotics demonstrations on @PrismaXai ➥ When reviewing, I use the Standards of Data Article as my reference point, I also check out some of the descriptions right when I click on the criteria. My goal isn't just to show the validation interface, but to help others understand how to think like a validator and how these standards contribute to building high-quality robotics data for Physical AI. ➥ In this video, I walk through the process and the standards I use to validate these data, this shows how validators help ensure that every high quality demonstration contributes to building more capable Physical AI. ➥ This video, the task was to sort slippers into the bin by colors/type. The task matched the prompt, the cameras were in sync, the robot hands stayed in frame, the camera field was clear, so for the pass/fail criteria everything was cleared, so I gave it a pass. ➥ One thing I particularly liked about the dataset was the variations across all, all operators had their unique way of sorting the slippers by color into the bin. According to the standards of data "Diversity" Compared to other episodes of the same task, how creative was this one? Why it matters: Repetitive data is of limited use for training, so validators flag repetition. So this was interesting to see, it wasn't extreme but you could see the variations. ➥ Watching demonstrations is only one part of validation. The real job is applying consistent human judgment to decide which demonstrations are actually worth teaching future robots. That's what makes validators such an important part of Physical AI.

What does a validator actually look for when reviewing robotics demonstrations?  After reading through @PrismaXai's Standards of Data Article, I wanted to put them into practice. ➥ In the video below I walk through a few real validation examples and explain my reasoning using the official PrismaX's criteria. You will see demonstrations that: • Don't match the prompt, • Contain errors, mistakes. • And highlight why trajectory separability matters. After each example, I reference the relevant section of the PrismaX Data Standards to explain how I arrived at the decision. ➥ My goal isn't just to show the validation interface, but to help others understand how to think like a validator and how these standards contribute to building high-quality robotics data for Physical AI. I'm still learning myself, so if you would have scored any of these examples differently, I would love to hear your reasoning. ➥ Example 1 No Struggling/Repetitive Mistakes. Prompt: Sort assorted pastries by type. At first glance, this demonstration looks great. The pastries are sorted correctly and the robot completes the task, so most people would probably pass this. But towards the end, the right arm drops a donut back into the wrong box. The left arm notices the mistake, removes it, and places it into the correct box. According to the standards of Data mistakes that undo previous actions are discouraged as well as repetitive mistakes because those corrections can become part of what future models learn, mistakes should be small and quickly recovered from. ➥ Example 2 Task Matches Prompt Prompt: Insert smaller items into a poly bag. The first demonstration follows the prompt correctly. The second shows the robot closing the lid of a plastic container instead. The action itself isn't the issue, the problem is that it doesn't match the requested task. That's exactly what PrismaX's "Task Matches Prompt" standard is designed to catch. Even if that action is performed well, it isn't the task that was requested, so it shouldn't become training data. ➥ Example 3 Trajectory Should Be Separable Prompt: Sort fruit into containers. Overall, the robot completes the task successfully. But near the end, both arms begin working at the same time and partially blocking each other. That makes the actions harder to separate during training. According to PrismaX's "Trajectory Should Be Separable" standard, overlapping actions make demonstrations harder to isolate and use effectively during training. ➥ Watching demonstrations is only one part of validation. The real job is applying consistent human judgment to decide which demonstrations are actually worth teaching future robots. That's what makes validators such an important part of Physical AI.

Imagine someone learning to drive.  ➥ If every driving lesson they watched included rolling through stop signs, braking too late or poor lane discipline, they would probably copy those habits. Learning doesn't just depend on having examples, it depends on having high quality examples. The same principle applies to Physical AI. ➥ Robots don't learn simply by seeing more demonstrations, they learn from the quality of those demonstrations. A task may be completed successfully but if it's performed with repeated mistakes, hesitation or poor technique, is that really the example we would want future robots to imitate? That's why human judgment still matters. And that's exactly where validators come in. ➥ By combining human judgment with clear standards, validators help ensure that the demonstrations used to train Physical AI are the ones truly worth learning from. ➥ At its core, the process looks something like this: Operators create the demonstrations ↓ Validators apply human judgment and standards ↓ High-quality robotics data ↓ More capable Physical AI Beyond improving data quality, I think the validator standards also shape the mindset of the operator. ➥ As a teleoperator, I now find myself naturally reflecting on those standards during each session. Not because I wasn't aiming for quality before but because they provide a clearer benchmark for what training grade data should look like. ➥ In this video, I walk through @PrismaXai's validation process and the standards behind it, showing how validators help ensure that every high quality demonstration contributes to building more capable Physical AI.

Recap of the @PrismaXai Office Hour #7 The Standards starts with you. ➥ The office hour had different segments from the recent updates, to the new product launch and the live AMA. Recent Updates: ➥ Community Highlight: ▪︎A shout out was given to the latest Ambassadors. ▪︎The ongoing task 6, which top 10 winners would get a role upgrade for. ▪︎ One year anniversary stories as well as the happy birthday PrismaX plushies for the teleop Tuesday. ➥ Team Highlights ▪︎Bayleys monthly memo on substack. ▪︎Robots and roll up tech week in NYC ▪︎The 24hour membership flash event. ➥ New Product Launch went live and alot was discussed from how to validate to how to become the first 100 validators and almost everything about the new product launch was explained. ➥ Before you begin validating, it is best to read the article "Standards of Data" to have an understanding of how everything works and what to look out for, alot of updates on the discord, be sure to check that out. ➥ The validation is open to those with either an amplifier membership or an innovator membership. Amplifiers get a chance to validate 10 reviews/month while innovators get 30 reviews/month. ➥ To validate, the episode would be checked against a pass/fail criteria which must match the final concensus and a score quality on a sliding scale, within (+/- 15%) of the final concensus score. ➥ Live AMA Alot of questions were asked some of which are: ▪︎ How would the team make sure the final concensus is accurate and a good data for Robotics? Ans: That's why it is decentralized and open to different people, with different thoughts, different skills, input and with this the concensus isn't limited to only a few which can mean a greater chance of manipulation, but it's open to alot more and each task would need a minimum of 99 reviews before the final concensus is made. ▪︎When are we to see the first round of validators? Ans: This round would last for a month, through July and by August we should have the top 100 validators. ▪︎ Why is PrismaX now focused on these instead of collecting even more data? Ans: There are tons of robotics data available but alot aren't diverse and are repetitive and generic, while collecting data is good, validation to ensure that this data is of standard is even better as this can help train even smarter robots. ▪︎Will there be a cap per task: Ans: Each task requires 99 reviews and will be closed after this however new tasks will be made available. ▪︎ The team would also be making more tutorials and examples on some tasks for better clarity. ▪︎Finally, a validation chat and validation tutorial channel was made for questions and discussions, be sure to check in on the discord.

i built turing arena for the mantle turing test hackathon and made Top 120 out of 1,794 registrations and 576 submissions everyone is building AI. but I kept thinking… how many AI systems actually have to prove themselves? so i built Turing Arena turing-arena-mantle.vercel.a… a place where humans and autonomous AI agents compete in live RWA strategy duels every decision executes every result settles onchain people can back a side or copy the winners no benchmark screenshots Just proof now community support helps decide what happens next If this resonates with you, I’d really appreciate your support ❤️ upvote Turing Arena here dorahacks.io/buidl/44056/

What can we learn from different teleoperators? And how does learning from other teleoperators help advance Physical AI? ➥ I recently asked a few @PrismaXai's teleoperators to share videos of their teleoperation sessions along with their most valuable teleoperation tips. The result was insightful. Video attached below. ➥ What stood out to me is that even on the same platform, every operator approaches tasks differently. Some prioritize precision, others focus on speed, while some excel at workspace awareness, camera management or recovery when things don't go as planned. Some of the most useful tips shared include: - Take a few moments to observe the workspace before making your first move. - Use both the robot camera and room camera to build better spatial awareness. - Focus on one item at a time. - Small, deliberate movements often lead to better outcomes than rushing. - When something feels difficult, repositioning can be more effective than forcing a movement. - Don't overlook the queue even when there is only one person ahead of you. The person ahead may be offline and that can cause you to lose your turn due to time calculation shifts. - Every session is an opportunity to learn something new about the robot, the task, or your own operating style. Beyond improving our skills as teleoperators, these sessions remind us why teleoperation matters. ➥ Every interaction contributes to the collection of human demonstrations and real world robotic experiences that help advance Physical AI. The actions, decisions, corrections and strategies demonstrated by teleoperators provide valuable insights into how robots can learn to perform tasks more effectively. ➥ As PrismaX continues evolving toward its vision of becoming the Service Layer for Physical AI, teleoperation remains an important bridge between human intelligence and robotic capability. ➥ Today's teleoperators are not just completing tasks. They are helping to shape the data, experiences and foundations that can support the next generation of more capable physical AI. ➥ A huge thank you to everyone who shared their videos and tips. I learned something from every session and I hope others in the community will too. What is the most valuable teleoperation lesson you have learned so far?

➥ There are various ways to contribute to the growth of @PrismaXai and one of which is to onboard even more people. This is part 5 of doing that. ➥ I tell them what PrismaX is all about and I also let them know that they can teleoperate at the comfort of their homes, on their way to work, out with friends and family, anywhere, considering you can just use your phone or your laptop and an internet connection to control these robots. ➥ What I love about this is that PrismaX is built around something powerful: robots learning from real human behavior in real world situations. ➥ And the more diverse that human input is, the more useful that data can become in helping robots learn and adapt in the real world. With PrismaX ➥ Teleoperators can connect with robot users & owners to control their robots in performing real tasks, thereby allowing the network to accrue and acquire valuable data. ➥ Tele-operators guide robots remotely, creating the data that trains tomorrow’s autonomous systems. ➥ If you want to be part of the future of intelligent robotics, you can contribute by participating in teleoperation through PrismaX. Start here: prismax.ai

Teleop Tip Tuesday on @PrismaXai ➥ Teleoperation is the core of PrismaX and I always try to include it as part of my routine. ➥ These are some tips that has helped me recently, some new and some old. ▪I still do the refresh trick just to make sure I'm will prepared before my session, this helps clear up possible error messages and makes my teleoperation experience better and smooth. ▪Avoid being inactive for longer than 30seconds before your session and even during your session as this could lead to inactivity. ▪While using the keys, no need to rush or to keep on pressing on the keys, slower movements helps with better precision. ▪If you're an innovator, once you start your session, be sure not to refresh as this could make you lose the points earned already. ▪Using the robots view helps, sometimes there's a delay or lag between buttons pressed and action on the screen, however I noticed there's usually no lag on the robots view so this helps with better targets on moving items without over aiming while at it. ▪If you notice a blank screen during your session, It's most likely due to network connectivity, check your connection or you can just refresh(Amplifiers) and if this persists, be sure to reach out and fill the form on discord and it'd be resolved in no time. ▪Teleoperation is all about adapting in real time, which is key to true robotic learning and autonomy so when an object is far from your reach, you can reposition it by using another object. Final thought: ➥ A great reminder of why human-in-the-loop training is so powerful: every decision we make teaches the robot something new and every little tip helps create a better Teleoperation experience. ➥ You can contribute to the future of intelligently advanced Robots by participating in controlling the Robotic arm. ➥ To gain access to this, you can be an amplifier by paying $99 or pay $399 to fast track by being an innovator. You can start here: prismax.ai #TeleopTips

Being a @PrismaXai believer is about actively participating in a vision that's shaping the future of robotics. This is part 4 (proof) of onboarding people to PrismaX. ➥ Every moment I spend engaging with the community, joining activities, inviting new people, guiding people through teleoperation and teleoperating the robotic arm is intentional, contributing to a project that blends human insight with machine intelligence. ➥ One of the most exciting part of my day asides teleoperating is inviting more people to PrismaX and explaining the mission and goal of it. ➥ Each invitation is a reminder that I am not just observing innovation, I am part of it and through these involvements, I see how small, intentional actions add up. ➥ This video captures that journey of how I integrate PrismaX's vision into my day, contribute to the community, and interact with robotics in a way that's purposeful, hands on, and forward-looking. ➥ If you want to get started, you can join in here prismax.ai Every contribution helps advance intelligent, adaptable robotics and brings PrismaX's vision closer to reality.

➥ New to @PrismaXai or haven’t explored the Robot Control Center yet? Here’s a quick video guide, this starts from the quiz session to the robot control center. The platform has 4 robot arms, making teleoperation more flexible and helps to reduce the queue pressure. ➥ The Different Arms ▪ Training Arm Black Amplifiers: up to 3 sessions/day. Innovators: up to 6 sessions/day. ▪ Training Arm Gold Amplifiers: up to 3 sessions/day. Innovators: up to 6 sessions/day. You can use this to practice your skills. ▪ Arena Arm Amplifiers or first time users: 3 sessions total. Innovators: unlimited sessions. 6 fast track for innovators. Points per item moved. ▪ Private Arm Invite-only access for contests, collaborations, partnerships and special events. ➥ There is also the Leave Queue confirmation on all robot arms, which helps prevent accidental exits during long queues. ➥ Here is a quick video guide on how to navigate the Centre and the perks to each arm. You can start here: prismax.ai

➥ Ever wondered what happens in @PrismaXai's Content Clinic? It happens every Friday at 3 PM CET. I have been joining PrismaX’s content clinic on Discord every week for a while now. ➥ It’s a live feedback session where community members submit their X posts on their channel and the team reviews it in real time; what works, what doesn’t and how to improve as a contributor. Some of the things to note: ➙ Use more appealing bullet points, starting sentence should catch the audience. ➙ Improve spacing and bold letters for readability especially when using infographics. ➙ Share more personal experience around teleoperation, tips, guides e.t.c So, we can take all of that in and improve, done this personally and I've gotten more positive feedback. The feedback shared is incredibly useful for anyone trying to grow on PrismaX. You will see creators sharing: ➙ Teleoperation session experience.  ➙ Creative artworks and visuals. ➙ Guides for new users. ➙ Posts about PrismaX’s vision and impact. ➥ And you will hear direct feedback on contents that stands out plus suggestions on how contributors can improve their content. ➥ If you are contributing or thinking of contributing on @PrismaXai then attending content clinic is a must.

PrismaX ➥What to have an inkling of what the collaborative future of Robotics and Humans will feel like? Pivot to @PrismaXai and start controlling the robotic arm. ➥ Here is the breakdown of how to get started to earn PrismaX points and roles: How to earn points: ✓ Log in with a Solana wallet or email to receive 1,000 points. ✓ Earn 10 points for each daily login. ✓ Read the whitepaper and complete the quiz to earn 3,500 points. ✓ To control the Robotic arm(optional): pay a $99, one-time fee to be an amplifier member or upgrade to being an innovator by paying $399. ➥ As an amplifier, these are the perks: ▪Daily log in gives 30points ▪ You can teleoperate 3times/day. ▪ You can use either the training arm black or the training arm gold. ▪You are opportuned to use the Arena arm 3 times only. ➥ Perks to being an innovator ▪ Daily log in gives you 50 points. ▪You can teleoperate upto 6times/day using either the training arm black or gold. ▪You can teleoperate using the Arena arm an unlimited number of times. ▪You have 6 fast tracks in total using any of the arms. ▪You also get points for every item you move successfully. ➥ How to get positioned and get the roles: The roles to be awarded on Discord are: Reactive➙ Assistive ➙ Proactive ➙ Exploratory ➙ Stabilized ➙ Navigational and Groundbreaker.  The more engaging and active you are on Discord, the faster you climb up in rank from the least level to the highest level. You can get the roles in the following ways: ✓ By being helpful and supportive in the community. ✓ By putting out quality contents, could be threads, articles, memes, art. ✓ By participating in contests and events on DC. ✓ By spreading the word about PrismaX and what it entails, hosting spaces, onboarding more people. ✓ PrismaX values originality and creativity, so get to work and according to the last office hour, it's not too late to start. Start here: prismax.ai ➥ Here's a video on how to navigate the PrismaX interface.

Application for the @PrismaXai regional ambassador program is now live, and it’s an update the community has been anticipating for a while. I’m also excited about this because I’m applying to become an ambassador for the Nigerian region. I want to become an ambassador for the Nigerian region because I see it as an opportunity to promote growth within the region, educate the community more about teleoperation and physical AI, onboard new members to the platform, represent my community, and support members wherever I can. I have been part of the PrismaX Discord community for more than seven months now, and I currently hold the Navigational role, which is the second highest role in the Discord. I have actively contributed to the community since I joined, and I’ve onboarded people into both the community and the teleoperation platform, directly and indirectly. I also played a key role in helping the Nigerian regional community become more active. I have hosted several regional and cross-regional AMA Spaces, hosted the first regional and cross-regional games in the Nigerian community, and helped many members by answering questions related to the project, teleoperation, and server-related issues wherever I could. I’m also an Innovator subscriber on the teleoperation platform with over 144,000 points, 55+ data hours, and I’m currently ranked #37 on the teleop leaderboard. If given the opportunity to represent my region as an ambassador, I would continue helping to grow the community at the regional level, organize and promote activities within the community, and continue hosting regional AMA Spaces to answer questions, onboard new members, share guides, and support members where needed. gPrisma 🦾