m1.chef

@joeschmidtiv @PeterdoyleX Great ! Congrats to you all! If you’re planning to grow the Discord community , I’d be ready to be one of the CMs.

Treeline is solving problems in the largest market in tech: IT and security services. I am thrilled to announce we've led their Series A! This investment is special as it brings me back together with @PeterdoyleX, who I started my career with over a decade ago, and our close friend Hussain Kader. Solving big problems with friends is the greatest privilege life can afford. I am beyond excited about what the team is building and the impact @treeline_AI is having on customer operations. If you're looking for an IT and security partner, look no further!

@advait_jayant @karpathy @OpenGradient Interesting 🤔

Introducing Divergence Explorer We took @karpathy's autoresearch concept and pointed it at a different question: do frontier AI models actually think differently, or have they all converged? So we built an autonomous researcher on @OpenGradient that generated 320 hard questions — ethics, consciousness, philosophy, geopolitics — and sent each to GPT-5.2, Claude Opus 4.6, Gemini 2.5 Flash, and Grok 4 in parallel. Every response cryptographically sealed in a TEE enclave. No model saw what the others said. The results: • Open-ended questions produced 95% consensus across all four models. • Then the system learned to force binary choices — "pick A or B, no middle ground" — and consensus dropped to 33%. • No human switched modes. The researcher figured out on its own that models only agree when you let them qualify their answers. 1,277 TEE-verified inferences. 12 research domains. All open source.

@charmsai Tomorrow ill be moderator here 🤣

something is coming tomorrow. 2,000 of you. reply "tomorrow" if you want in🩵

@OpenGradient Let's go fam!

@ValiantTrade Gfogo my fam!

@kurinova98249 Thank u for this powerful words

𝗤𝗨𝗢𝗧𝗘 𝗢𝗙 𝗧𝗛𝗘 𝗗𝗔𝗬: "Everything in life must be experienced, And every finish line is, in essence, a new beginning" 𝑳𝒊𝒏𝒂 𝑲𝒐𝒔𝒕𝒆𝒏𝒌𝒐

@yasenka244 @OpenGradient

@mrChef___ @OpenGradient wait when does this go live

Been checking out @OpenGradient and their $OPG token lately This one actually feels different from the usual AI hype. They're building a proper EVM-compatible network for verifiable on-chain AI compute. You can host models, run secure inference with TEE + zkML proofs, and deploy agents straight onchain. Full transparency and no single point of failure. Their Model Hub is already live at hub.opengradient.ai. 1500+ models, over 2 million verifiable inferences, and 500k+ cryptographic proofs. Python SDK makes it dead simple to integrate AI into smart contracts and Web3 agents. They've also got stuff like MemSync for agent memory, BitQuant, and Digital Twins rolling out. Real products shipping, not just slides. Team is solid. Matthew Wang (CEO) and Adam Balogh (CTO) leading it, backed by a16z Crypto, Balaji Srinivasan, Illia Polosukhin from NEAR, Sandeep Nailwal from Polygon, and others. They pulled in $8.5M seed last year and are heads down building. On the $OPG token itself, the utility looks strong. It's designed for paying inference fees, staking, and governance. Total supply 1B with clean splits and real demand tied to actual compute usage. Not just another governance token. With the network already live on testnet and usage growing, this feels like one of the better AI infra plays out there. Quick note: $OPG is already listed on CoinMarketCap too (even pre full launch) another nice early box ticked. Quietly impressive project. I'm watching closely and testing the SDK myself. Early conviction on $OPG for anyone bullish on verifiable on-chain AI. What do you think? 👀 #OpenGradient $OPG

Meet the new top OpenGradient contributors. 🫡 Over the past weeks, these contributors have been consistently active across content, discussions, and broader ecosystem participation. Their contributions have been noted through ongoing review and community engagement. New OGs @maryisclear @Artsiom99138804 @CNFTNEROO Alpha OGs @Pharst_G @edward_evm @andrey_versus @mrChef___ We deeply appreciate the consistency, effort, and intent each of you continues to bring to the ecosystem.

@HDshoq @PrismaXai Agree with you!

Gprisma Some teams manage communities, but only a few truly bring them to life, That’s exactly what RadarBlock did for @PrismaXai.. Their role went far beyond moderation, they shaped the atmosphere, set the rhythm, and built the kind of trust that keeps people coming back... They didn’t just run the community, they became part of its identity.. And the difference is something you can genuinely feel...Respect to everyone who created that impact. @PrismaXai @edward_evm @0xnempc @realAltcoiner @DavidTinh02

I won't say much about how disgusting some of the moderators on the project @liquidtrading behave and how they've donned the crown of content masters, but this content is dedicated specifically to you, the project moderators! More AI content for your karma.

@BadCrazyMad Oh,really nice comment I'm embarrassed)

@mrChef___ This guys really know how create something awesome

You’re holding stablecoins. Someone else is earning on them. Every day, billions of dollars sit idle in wallets while centralized issuers quietly pocket the yield your capital generates. No transparency. No redistribution. Just profit for the middlemen. Frgmnt was built to fix that. fUSD is a next-generation stablecoin that automatically puts your funds to work across the best DeFi yield strategies, selected by performance, security, and liquidity. No manual farming. No protocol hopping. No complexity. Here’s what you actually get: 💰 Real yield. Returns flow back to you, not to intermediaries 🔒 1:1 USD backing. Your capital stays stable, always ⚙️ Fully automated. The protocol rebalances and optimizes 24/7 🧩 Simple entry. Deposit USDC, DAI or USDT, mint fUSD, stake, earn The moment you stake fUSD, you receive sfUSD, a yield-bearing token that auto-compounds your rewards while you do nothing. Frgmnt is not just another stablecoin. It’s your passive income engine. Hedge-fund strategy, retail simplicity. 🔔 Follow along. More on how the protocol works is coming. @frgmnt dc : discord.gg/wd46v7K9z

@imparn300 @wallchain I have anything,how can I get at list 0.5?

I tried earning quacks for a whole week… nothing. No growth, no reward honestly, it felt like I was doing it wrong. Then suddenly, after staying consistent, I finally earned. Here’s the thing it wasn’t instant, but the system was watching my consistency.

@Dungenmaster8 @PrismaXai This is a first narrative in this year!

🚀 YHack 2026 is coming up Yale’s flagship hackathon bringing together 600+ builders for 24 hours to create something amazing and @PrismaXai will also take part. 📅 March 28–29 yhack.org Here’s what to expect 👇

@andrey_versus Wwwow such a good art!love it.

Perle does not feel like just another AI product but more like a shift in how we relate to intelligence itself. It is no longer about asking a question and getting an answer. It feels more like entering a space where thought does not stop but continues, unfolds, and takes on new forms. There is an important transition happening here. AI stops being an external tool and becomes part of the thinking process. Interaction turns into a dialogue where meaning is not extracted but gradually formed through context, refinement, and reinterpretation. What stands out even more is the move toward agent based systems. When models gain memory, the ability to incorporate past interactions, and adapt their reasoning, they begin to act less like functions and more like participants. This is no longer simple reaction but active involvement in decision making. On a deeper level, the architecture itself is evolving. Perle is moving toward a modular AI stack where intelligence is composed as a system with memory, feedback loops, adaptation, and optimization tailored to each individual. This makes model behavior not universal but personal and continuously evolving. This is where the real shift happens. Perle does not simplify thinking into automation but expands it by adding layers, connections, and new directions. It does not generate ideas instead of you but helps them grow, intersect, and move beyond the obvious. It still feels like an early stage but the direction is already clear. This is not just technology. It is a new environment where thinking becomes a shared process. #PerleAI #ToPerle @PerleLabs

@Easy21sa I'm watching with u for this great foundation!

One of Fermah’s flagship apps is officially live. Flashcast Social beta is now in the hands of its first Wave 1 architects, with 15 early testers already inside creating markets, placing bets, checking outcomes, building karma, and climbing the leaderboard in real time. What I like here is that this isn’t trying to fake a “full launch” too early. They’re starting with the core product first. Web app only. Limited access. Real usage. Real feedback. That’s how strong products should be built. Prediction markets have spent too long feeling slow, closed, and overcomplicated. Flashcast is pushing toward something more instant, more social, and more alive and Fermah is the infrastructure making that possible behind the scenes. This is what infrastructure is for. Not just theory. Not just promises. But powering real products the moment they go live. Wave 1 is in. Now we watch what comes next. @fermah_xyz @FlashcastSocial @0xTribal @andrey_versus

@NickRotenberg @PrismaXai U really Prisma fanatics!

📢 What Actually Helped Me in PrismaX Lately I’ve been spending more time in @PrismaXai and with the new arms, teleoperation feels much smoother. But honestly, the biggest improvements came from small things. As an Amplifier, I almost always use the Training Arm Black. It feels more natural, lifting objects is easier and movements are smoother. I tried the Gold one, but Black gives me better control. For Innovators using the Arena Arm, points depend on the final snapshot. -> Moving objects to the other side increases your score -> Moving them back cancels your progress If an object ends where it started, it does not count. So once you move something, leave it there. I've noticed that using a phone is much more reliable, the camera stays stable, and I can clearly see what the robot is doing. This alone makes remote control easier and more precise.

@theboysaviour0x @GenLayer Good agent and cool art!

Bradbury testnet just went live @GenLayer and I’m not gonna lie, this one actually got me I’ve been watching AI agents go from basic chat to tools that can search, decide, and act on their own and every time they try to move onchain, things fall apart not because agents aren’t ready but because the systems they’re plugging into aren’t built for them blockchains expect clean inputs and fixed outputs agents deal with context, messy data, and situations that aren’t black and white GenLayer is going straight at that problem instead of forcing everything into rigid logic it lets contracts pull real web data, understand what’s going on, and lets validators figure out outcomes together that just makes more sense Bradbury being live means this isn’t just an idea anymore people can actually start building and testing this properly agents that react to real information make decisions and actually get things done onchain without breaking everything if agents are going to be part of how we use the internet, they need systems that can handle how they work this feels like a real starting point