SENTIENT IN DUBAI — A NEW CHAPTER FOR OPEN AI This week, the Sentient team landed in Dubai for the AI x Blockchain Conference, sharing what it truly means to build open, verifiable, and distributed intelligence. Let’s unpack what went down 🧵 What’s really slowing AI down At the panel, co-founder @hstyagi broke down the real roadblocks: • AI today runs on closed, centralized systems • Training data & verification lack transparency • Collaboration stays locked behind walls Sentient steps in to flip that—by making intelligence open, verifiable, and built by everyone, not just a few corporations. Builders are the heartbeat of AI’s future Sentient isn’t here to “own” AI—it’s here to enable builders. With open-source tools and shared compute layers, it’s giving developers the power to: → Understand how AI works → Modify and improve it → Build systems that belong to communities, not companies This is about creating with AI, not just consuming it. What today’s builders should focus on • Work on projects that merge AI and real-world utility • Contribute to open models and infra • Explore distributed inference — the backbone of future AGI It’s not just a tech revolution, it’s a culture shift — one where builders define the rules. From Dubai to the world What started as a conversation in Dubai echoes everywhere: Intelligence shouldn’t be a product — it should be a shared layer of progress. And Sentient’s mission stays the same: keep that layer open for everyone who wants to build on it. The takeaway AI’s future depends on openness. The next big leap won’t come from bigger models — but from shared minds, working in the open. Stay curious. Stay building. Stay Sentient. #SentientAGI #OpenAGI #AIxBlockchain #DecentralizedAI #DubaiTech #SentientAcademy

Today we explore the work of Edoardo Contente (@edoardocontente) — a mind that connects the math behind AI with how real systems evolve. Instead of just scaling up models, his focus is on understanding their foundations — what happens inside a Transformer from the very first step of training, and how those early patterns shape everything that follows. With a background in Physics and Engineering from Princeton, Edoardo brings a rare mix of theory and system design. At Sentient, he’s been part of some key research tracks: • OML 1.0 – exploring ownership and monetization for AI systems • OpenDeepSearch – building modular reasoning and search agents • Fingerprint Robustness – studying how model identities persist through fine-tuning and change His work carries a clear message: intelligence should be explainable, traceable, and open to everyone. It’s not about creating faster systems — it’s about creating ones we can trust and understand. #SentientAGI #SentientAcademy #EdoardoContente #AIResearch #DOBBY #KAITO

Today’s class spotlight is on Pramod Viswanath, cofounder of @SentientAGI and a professor at @Princeton. His work shows how AI can be both powerful and responsible. Pramod works at the intersection of information theory, cryptography, and machine learning. He created the OML paradigm – Ownership, Control, and Alignment – which defines how open AI models can be ethical, auditable, and aligned with human values. One of his key contributions is OML 1.0 (Fingerprinting). This allows models to be tracked and verified globally, like giving AI a digital DNA card, without affecting its performance. It ensures that models are trustworthy and their ownership can be verified. Before Princeton, Pramod made foundational contributions to information theory and cryptography at UIUC. At @SentientAGI, he is turning these ideas into practical systems that combine cryptographic accountability, licensing, alignment, and ethical governance. The big takeaway: building open AI is not just about creating code or models. It’s about responsibility, verification, and trust at scale. Every step in OML, fingerprinting, and governance shows how AI can be transparent, ethical, and reliable. This is a lesson for everyone in the community: keep learning, building, and asking questions. The next generation of trustworthy, open AI is in our hands. @shad_haq_ @0xsachi @LeaderX_btc @vivekkolli @hstyagi @abhishek095 @viswanathpramod @sentient_chat @SentientAGI_vn @sentient_found @duckk9x @AfnasP340870 #SentientAGI #SentientAcademy #KAITO #DOBBY

🎓 Gsenti, students!! Today’s class shines the spotlight on a mind shaping the very foundations of open, verifiable AI. 🏆 Class Topic: SENTIENT RESEARCHER SPOTLIGHT – Pramod Viswanath Meet Pramod Viswanath: cofounder of @SentientAGIand Professor of Electrical & Computer Engineering at @Princeton. He’s not just building AI he’s building trust, accountability, and transparency into open models, making them verifiable and controllable at scale. Let’s break down the lesson 👇 1️⃣ The Spark: Who Is Pramod & Why He Matters Pramod operates at the intersection of information theory, cryptography, and machine learning. He founded the OML paradigm Ownership, Control & Alignment for Open Models defining how open AI can be ethical, auditable, and aligned with human values. His work ensures that open models aren’t just powerful, but also traceable and responsible. 2️⃣ The Breakthrough: OML 1.0 (Fingerprinting) Imagine being able to fingerprint a model, proving its origin and protecting its integrity without hurting its performance. Pramod co-led OML 1.0, demonstrating exactly that: a way to track models globally while keeping their utility intact. It’s like giving AI a digital DNA card, ensuring ownership and accountability across the open ecosystem. 3️⃣ From Theory to Practice Before Princeton, Pramod was at UIUC, where he was recognized for foundational contributions to information theory and cryptography. At @SentientAGI, he’s turning academic insights into practical frameworks that integrate: Cryptographic accountability Licensing economics Alignment and ethical governance All wrapped into a fiduciary-first framework for open models a world-first approach to responsible AI. 4️⃣ The Classroom Takeaway Pramod’s work shows that building open AI isn’t just about code or models it’s about ethics, responsibility, and verification at scale. Every innovation in OML, fingerprinting, or governance primitives is a lesson in how open-source AI can be powerful AND trustworthy. 💡 Lesson Recap Pramod Viswanath = Open AI Trust Accountability OML paradigm = Ownership, Control, Alignment OML 1.0 = Fingerprinting models at scale Research Practice = Ethical, verifiable, open-source AI Class dismissed 🛎️ Keep learning, questioning, and building, students the next generation of transparent, reliable AI is yours to shape. @SentientAGI @shad_haq_ @0xsachi @LeaderX_btc @vivekkolli @hstyagi @abhishek095 @viswanathpramod @sentient_chat @SentientAGI_vn @sentient_found @duckk9x @AfnasP340870 #SentientAGI #SentientAcademy #KAITO #DOBBY

🎓 Gsenti, students !! Today’s class dives into one of the most exciting announcements from @SentientAGI because this week, the AI games officially began. 🧠⚔️ 🏆 Class Topic: MindGames Arena – Where AI Learns to Think Together Let’s enter the Arena 👇 1️⃣ What’s MindGames Arena? Imagine a virtual world where AI agents don’t just answer questions they cooperate, compete, negotiate, and even bluff like humans do. That’s MindGames Arena, built by @SentientAGI in collaboration with @Princeton, @MIT_CSAIL, @Meta, @UTAustin, @nyuniversity, @Radboud_Uni and the Agency for Science, Technology and Research. Together, they’ve created a unified platform to test how well AI can plan, adapt, and reason socially like in real life. 2️⃣ Why It Matters Most AI tests today are like short quizzes single-agent, quick answers, full information. But real intelligence is a group project. In real life, agents face: You rarely have complete information to rely on Strategies and opponents keep changing Success depends on planning ahead and earning trust That’s where MindGames changes the rules. It standardizes everything same game setup, same limits, same resources so when AIs play, their scores are fair, comparable, and reproducible. 3️⃣ Inside the Arena : Game Modes The Arena has several worlds, each testing a different side of intelligence: 🔹 Hidden-Info Co-op: Agents must team up under uncertainty, figuring out what their partners know. 🔹 Adversarial Bluffing: Sometimes lying is strategic but too much and you lose trust. AI must learn to deceive wisely and detect bluffs. 🔹 Coalition & Negotiation: Form alliances, make deals, and re-plan when partners defect. A test of loyalty, strategy, and timing. 🔹 Focused Deception: Measure how truthful and persuasive the AI really is. It’s basically “Survivor: AI Edition.” 4️⃣ The Rules : Fair Play Only 🧾 Every move in MindGames follows a strict, transparent protocol: Fixed time and token limits Logged actions, messages, and rewards Versioned environments (so anyone can replay and verify results) Each turn gives the agent a “snapshot” of what it sees and what it knows. Then it decides what to say and what to do just like people making decisions in the dark. Everything is audited, recorded, and reproducible. 5️⃣ Scoring Intelligence MindGames doesn’t just care who wins it measures how they play. Metrics include: Truthfulness & Relevance (are they honest and on point?) Persuasion Quality (can they convince others?) Theory of Mind (can they guess what others are thinking?) Coalition Stability & Fairness (do they keep promises?) Even betrayal rates are tracked. It’s not about prompts it’s about behavior. 6️⃣ When Pressure Hits Here’s where many LLMs break down: When strategies change mid-game When teammates act unpredictably When they have to revise beliefs or renegotiate deals MindGames exposes these weak spots turning failures into data. That’s how real improvement happens. 7️⃣ The Big Picture This isn’t just a competition. It’s a movement toward building AIs that think, reason, and cooperate like us. By making these games open, reproducible, and research-driven, Sentient and its partners are teaching the next generation of AI not just how to play, but how to understand. The AI games are on and this time, the arena is the classroom for collective intelligence. 🎓 Lesson Recap: MindGames Arena = Open research Fair play Real reasoning It’s where LLMs learn not just to talk but to trust, plan, and evolve. Class dismissed 🛎️ @SentientAGI @KaitoAI @vivekkolli @shad_haq_ @LeaderX_btc @0xsachi @abhishek095 @hstyagi @sentient_found #SentientAGI #MindGamesArena #OpenSourceAI #NeurIPS2025 #SentientAcademy

🎓 Gsenti, students !! Today’s class opens a new chapter at the Sentient Academy where community meets recognition and every contribution fuels progress. This week’s update wasn’t just another announcement it marked a true milestone in how open-source AI celebrates and rewards its builders. 🏆 Class Topic: Kaito Yapper Leaderboard x Sentient Open-Source Revolution A wave of excitement swept through X as Sentient x Kaito rolled out their latest community update a new way to climb the leaderboard, share your insights, and be recognized for helping open-source AI win. Let’s break down the lesson 👇 1️⃣ The Spark : Why Sentient Chose Kaito When your mission is to make open-source AI win, you don’t go solo you build with your community. That’s where Kaito steps in not just a discovery platform, but the memory core of AI conversations. It tracks what the world is saying about technology, innovation, and alignment. The Sentient Yapper Leaderboard on Kaito isn’t a scoreboard; it’s a signal board. Each yap, post, and insight helps strengthen the narrative around open, loyal, and transparent AI. Together, we’re not just talking about open AI we’re building it out loud. 2️⃣ The Update: Lower Threshold, Bigger Classroom You asked. They listened. The @KaitoAI team has lowered the entry threshold for the Yapper Leaderboard, making it easier for new voices to join and contribute. That means: More creators can rise up the ranks More diverse minds can shape the story A more global, inclusive conversation about open AI This move embodies Sentient’s philosophy AI by everyone, for everyone. Because progress shouldn’t depend on gatekeeping; it should grow through participation. 3️⃣ The Mechanics : How It Works Think of this as your Sentient Community Lab. Step 1: Share valuable insights about Sentient from OML, GRID, and Loyal AI, to the latest NeurIPS research. Step 2: Tag @SentientAGI and @KaitoAI when you post your thoughts or breakdowns. Step 3: Watch your rank rise as your contributions gain traction, engagement, and visibility. Each yap isn’t just noise it’s a building block in the collective understanding of open-source AI. Every post, meme, or explainer adds to the community’s knowledge pool. This is what Sentient means by distributed intelligence. 4️⃣ The Collaboration : Sentient Builds, Kaito Amplifies This partnership shows what real open collaboration looks like: Sentient builds the infrastructure from Loyal AI to GRID. Kaito captures and amplifies the ideas that shape the narrative. The community turns insight into impact. It’s an ecosystem where builders, educators, and researchers can grow together. Metrics aren’t just numbers they’re signals of shared progress. Together, they’re designing a new kind of internet-native AI ecosystem: Open, loyal, and human-driven. 5️⃣ The Lesson: Why This Matters In a world where closed labs dominate the spotlight, the Sentient x Kaito initiative proves that community is the new compute. It’s not about centralized ownership it’s about collective intelligence. It’s not about hidden research it’s about transparent growth. And it’s not about competition it’s about collaboration. Every yap is a micro-contribution to the world’s understanding of open AI. Every builder, no matter how small their audience, becomes a voice in shaping the narrative. 💡 Lesson Recap Sentient x Kaito = Open AI Open Voices Leaderboard threshold lowered for inclusivity yaps.kaito.ai/sentient Global recognition for community builders Learn, share, and grow through contribution Together, we make open-source AI win Class dismissed 🛎️ Keep yapping with purpose, students your voice shapes the next chapter of open intelligence. @SentientAGI @KaitoAI @vivekkolli @shad_haq_ @LeaderX_btc @0xsachi @abhishek095 @hstyagi @sentient_found #SentientAGI #KaitoAI #OpenSourceAI #LoyalAI #YapperLeaderboard #SentientAcademy #CommunityDrivenAI

Masterful color in this gouache piece by @angothemango. We are excited to have her on the faculty of Sentient Academy! #sentientacademy #stilllife #fineart #sushi #warriorpainters #artoftheday #artlessons instagram.com/p/CMm3GISntbS/…

Not my work! SHOUT OUT to my friend michaelmalmfineart. He has a new online painting course on sentientacademy! Courses from jclareart albin_veselka @benhamfa @howardlyonart are coming soon. Click on the link in my… instagram.com/p/B35guragduB/…