NFTs were an open book. Every trait, every rank, scraped before the reveal even finished. @FungoLabs sealed them. First encrypted NFTs on Ethereum. Built with @zama FHE. Every mint is born encrypted onchain. The contract works on data it can never read. Art is yours to keep sealed or reveal. The essence stays hidden forever only the holder sees it. Seen by all. Known by one. Whitelist is open: fungolabs.org/apply/

you’ll never hear people comparing prediction markets to onchain again Even to see big p&ls of onchain traders u had to be in a niche algorithm before Permanent onchain employment there will never be another “bear market” where there’s ZERO money to be made again

$5M mcap today?

I've been thinking a lot about where AI is actually headed For a while, every conversation was about which model is bigger, faster, or smarter. That stuff matters, but the more I use AI day to day, the more I realize the real question is different: does it actually make my work easier? Most of us don't need a more powerful model. We need something that fits into how we already work, create, and think. Something that takes a messy, complicated task and makes it feel simple. That's where the real value is. This is why I'm keeping an eye on @lynote_ai . They seem to be building around real human needs instead of chasing hype, and I like that direction. It's still early for this space, and I think we've only seen the beginning

WHAT IF THE NEXT BREAKTHROUGH IN WEB3 ISN’T ANOTHER APP, BUT THE INFRASTRUCTURE BEHIND IT? That's where @NucleusCodes gets interesting Nucleus is focused on building a developer-first, composable ecosystem designed to make it easier to turn ambitious ideas into real products. Less friction. More flexibility. More room for builders to experiment and scale. The infrastructure layer is often overlooked until it becomes impossible to ignore.

Most people will watch Papertrade launch PaperDAO wants to be positioned for the first block. @paperdao_xyz starts with a $0 LP, no seed LP, no pre-mint, no team/VC PAPER. PaperDAO pools community capital into one treasury so we can trade the front of the curve together. PULP represents our share of the treasury. Apply now: daos.world/launch/paperdao/a…

Still waiting for my y @gomo_family reward. It's already been more than 6 hours, even though I was told it would be sent within a few minutes. I’m not sure why there’s such a delay Hopefully, I'll receive my reward soon

gomo is the social Solana terminal: trade any coin, post your thesis, follow the chads. Go or miss out. Trading it on @gomo_family 🟦 the social Solana trading terminal CA: 9XKzy4KahcZaGJPJtz1PtqGPB3CiseoBrx7TcQhEpump gomofamily.life

The best plays usually look before they look obhious. FungoLabs is one I’m watching early @FungoLabs is building quietly while most of CT is busy chasing the next shiny narrative No forced hype Just product, experiment and an ecosystem that could get interesting with time Early is where the real asymmetric bets live keeping this one on the radar Secure spot: fungolabs.org/apply/

How we use Zama ? → Every NFT mints sealed and is encrypted onchain using Zama’s FHE. → The contract computes on data it can never read. → Its look and essence are determined by FHE randomness. Nobody gets to choose. Not even us. 🟨 Built with FHE on @zama.

Axis robotics is tackling one of the biggest bottlenecks in Physical AI. The problem is not only better robot hardware or larger models. It is high quality training data. Robot learning depends on state action trajectories, and the physical state space is enormous. Small changes in object pose, friction, lighting, contact, or robot configuration can produce completely different outcomes. @axisrobotics is building a data engine around this problem. The system combines procedural simulation, teleoperation, autonomous policy execution, and human intervention to continuously generate training trajectories. The interesting part is Human Gated DAgger. Instead od humans controling evry step, the policy attempts the task autonomously and humans intervene when it reaches critical failure states. Those corrections become new training data. This creates a feedback loop. That means the dataset can increasingly focus on the models failure distribution rather than simply becoming larger. This is the bigger thesis behind Axis. The future of Physical AI may not depend on who collects the most data. It may depend on who can continuously identify what the model does not know and generate exactly that experience at scale. That turns robotics data from a static dataset into a compounding intelligence engine. s.kaito.ai/9kNz0ol

What if a bunch of wallet could trade like one big wallet from day one? That is the idea behind @paperdao_xyz. Trading alone has its limits. Small wallets often cannot get good entries, cannot take meaningful positions, and have to carry all the risk by themselves. PaperDAO is trying to change that by building a community treasury where capital is pooled, so participants can trade together at meaningful size instead of each going it alone. The key piece here is $PULP. Holding $PULP gives you a pro-rata claim on everything the treasury holds, including trading capital, PAPER, staking yield, and other treasury assets. So your share grows or shrinks along with the treasury itself. The treasury is being raised through daos, with plans to move to HyperEVM and deploy capital around launch. That means the capital is not meant to sit idle. It is meant to be put to work from the start. The model is simple. Community capital goes into pooled trading, and pooled trading aims to create treasury upside that the whole community shares in. Of course, this is trading capital and not guaranteed returns. Markets can go against you, and $PULP can go to zero. Always do your own research and never put in more than you can afford to lose. Stil a community owoned treasury like this is definitely worth keeping an eye on. Apply link: daos.world/launch/paperdao/a…

Physical AI's real bottleneck isn't compute. It's data LLMs scaled on trillions of text tokens that already lived on the internet. Physical AI has no such luxury. There is no ready-made library of robot motion data. To learn how to grasp, move, place, and recover when things go wrong, a robot needs massive amounts of real and simulated trajectory data. The old way, small lab teams collecting data one machine at a time, is expensive, slow, and nowhere near the pace of model iteration. as @axisrobotics founder chris puts it:"Physical AI demands billions of human-physical interaction motion trajectories." Axis's answer: put the simulator in the browser. Anyone can open a webpage, operate a virtual robot,and complete a task. Every action becomes a trajectory sample. It's less a robotics company, more a distributed data production network: -100,000+ active contributors -1,200+ hours of simulation data every month -20,000+ hours of real-world first-person data every month The most interesting part is Axis V2's human-in-the-loop correction cycle. The model attempts the task on its own first. Humans step in only when it fails. Contributors no longer re-teach what the robot already knows. They supply high-value data exactly where the model gets stuck, shifting collection from accumulating volume to generating targeted quality. The benchmark results back this up. On LIBERO-Plus, pretraining π0.5 on Axis's diversified dataset improved success rates by 4.9 percentage points. It also beat a volume-matched RoboCasa365 baseline by 31.3 points. The lesson isn't "more data" but more diverse data, covering layout generalization, sensor noise, and robot pose robustness. Where this leads: As robot hardware gets cheaper, the moat moves from hardware to a single question: who can continuously produce high-quality, task-aligned training data? Axis is betting that the data engine itself becomes the infrastructure of the Physical AI era. They aren't building robots. They're building the fuel that makes robots smarter. Go here : s.kaito.ai/9kNz0ol

Maybe something Big is coming for @world_xyz users If you still don’t have an account on world, this might be your last chance to get in early. Something exciting could be waiting just around the corner👇 🔗world.xyz/?ref=6KKADZBE

Gm @axisrobotics Been thinking about this a lot lately. Most of us spend our whole week trying to squeeze more out of 24 hours. Robots don't have that problem. They don't get tired, don't take sick days, don't have a bad Monday. They just keep running. Think about what that actually means. A factory that runs around the clock with almost no downtime isn't just a little faster, it's a completely different kind of output. Fewer errors, steadier quality, and it scales without everything falling apart. The old model was built around human limits: shifts, breaks, fatigue, mistakes. None of that is anyone's fault, it's just how manual work goes. But the gap between that model and what automation can do is getting wider every year, and it doesn't go back. That's why @axisrobotics caught my attention. It doesn't feel like a "someday" idea. the work is happening now, on the floor, day after day, while most people are still debating whether it's real. I'm not saying it's a sure thing, nothing is. But i'd rather pay attention early than write "i should've looked into it" later. So, early or watching from the sidelines? your call. gm. let the robots cook 🤖 check: s.kaito.ai/9kNz0ol

Build. Create. Get recognized. Action Model is looking for creators and ambassadors to help shape the AI x Web3 movement. No massive following needed. Just ideas, consistency, and execution. Application are open

The X Premium gifting era is over 🥲 No more random premium gifts No more surprise 3 to 6 month gifts

Updated my profile picture picture today. A new look, but the same commitment to quality, learning, and meaningful collaboration. Grateful for everyone who has supported me along the way