Prey 🍊

𝐑𝐎𝐌𝐀 (𝐑𝐞𝐜𝐮𝐫𝐬𝐢𝐯𝐞 𝐎𝐩𝐞𝐧 𝐌𝐞𝐭𝐚 𝐀𝐠𝐞𝐧𝐭𝐬) 𝐯𝟎.𝟐.𝟎 has been released, completely rebuilt on DSPyOSS, this version brings human-like recursive reasoning to multi-agent systems in a practical and scalable way. 𝗞𝗲𝘆 𝗳𝗲𝗮𝘁𝘂𝗿𝗲𝘀 𝗼𝗳 𝗥𝗢𝗠𝗔: ❃…

𝐓𝐚𝐥𝐮𝐬 𝐱 𝐒𝐞𝐧𝐭𝐢𝐞𝐧𝐭: Expanding the frontier of decentralized AI. @Talus_Labs is partnering with @SentientAGI, the community-driven AI platform and research body behind AgentHub, SentientChat, and GRID, to bring onchain AI agents to millions of users and unlock real…

Meet 𝐑𝐎𝐌𝐀 𝐯𝟎.𝟐.𝟎, the next evolution of recursive, high-performance multi-agent systems. Rebuilt from the ground up on DSPyOSS, 𝐑𝐎𝐌𝐀 is designed to mirror human reasoning through recursive hierarchical decomposition, breaking complex tasks into smaller, independent…

𝐑𝐞𝐜𝐮𝐫𝐬𝐢𝐯𝐞 𝐈𝐧𝐭𝐞𝐥𝐥𝐢𝐠𝐞𝐧𝐜𝐞 𝐣𝐮𝐬𝐭 𝐞𝐯𝐨𝐥𝐯𝐞𝐝. 𝗥𝗢𝗠𝗔 (𝗥𝗲𝗰𝘂𝗿𝘀𝗶𝘃𝗲 𝗢𝗽𝗲𝗻 𝗠𝗲𝘁𝗮 𝗔𝗴𝗲𝗻𝘁𝘀) 𝘃𝟬.𝟮.𝟬 marks a major leap in multi-agent reasoning, rebuilt from the ground up on DSPyOSS for reliability, scalability, and true recursive…

𝐑𝐎𝐌𝐀 v0.2.0 is here. Recursive Open Meta Agents have been completely rebuilt from the ground up on top of the DSPy framework, faster, more reliable, and ready for real-world, production-scale recursive multi-agent systems. At its core, 𝐑𝐎𝐌𝐀 mirrors how humans solve…

𝐑𝐎𝐌𝐀 (𝗥𝗲𝗰𝘂𝗿𝘀𝗶𝘃𝗲 𝗢𝗽𝗲𝗻 𝗠𝗲𝘁𝗮 𝗔𝗴𝗲𝗻𝘁𝘀) 𝘃𝟬.𝟮.𝟬 𝗶𝘀 𝗵𝗲𝗿𝗲! The new release completely rebuilds the framework on top of DSPy, bringing a faster, more reliable, and builder-friendly ecosystem for high-performance recursive multi-agent systems. At its…

𝐑𝐎𝐌𝐀(𝗥𝗲𝗰𝘂𝗿𝘀𝗶𝘃𝗲 𝗢𝗽𝗲𝗻 𝗠𝗲𝘁𝗮 𝗔𝗴𝗲𝗻𝘁𝘀) v0.2.0 is here, a high-performance, production-ready recursive multi-agent framework fully built on DSPy. ➢𝗖𝗼𝗿𝗲 𝗠𝗲𝗰𝗵𝗮𝗻𝗶𝘀𝗺: When a task arrives, 𝗥𝗢𝗠𝗔 first asks: “Can this be solved directly?” ➫ If…

𝐑𝐎𝐌𝐀 (Recursive Open Meta Agents) v0.2.0 is here! 𝗦𝗲𝗻𝘁𝗶𝗲𝗻𝘁 has rebuilt multi-agent reasoning from the ground up using DSPyOSS, creating a builder-friendly ecosystem for high-performance recursive workflows. ROMA mirrors human problem-solving with recursive…

Sentient dominates @NeurIPSConf 2025 4 research papers. 1 mission, building the future of open, verifiable AGI. ★ 𝗢𝗠𝗟 𝟭.𝟬 (Main Track) A breakthrough in LLM fingerprinting, 100× scale-up, 24,576 persistent prints, zero performance loss. A new security backbone for…

✦⋆✦⋆✦⋆✦⋆✦⋆✦⋆✦⋆✦⋆ S E N T I E N T ✨ The open frontier of intelligence. 💠 OML 1.0: fingerprinting that protects every open model. 💠 LiveCodeBenchPro: efficiency as art, not compromise. 💠 MindGames Arena: agents that think, play, and evolve. 💠…

Fortytwo’s Swarm Inference architecture is challenging the dominance of centralized AI models. Instead of relying on a single giant model, it’s built on a distributed network of hundreds of nodes working together as one intelligence. The results are remarkable: ✶ ChatGPT 5, ✶…

𝗦𝗲𝗻𝘁𝗶𝗲𝗻𝘁 𝗔𝗚𝗜 has achieved a major milestone at 𝗡𝗲𝘂𝗿𝗜𝗣𝗦 2025, with four groundbreaking projects accepted that are redefining open-source AI. These works make significant strides in security, efficiency, and adaptive learning, offering tangible opportunities for…

Every truly open system eventually learns a paradoxical truth: privacy isn’t the enemy of transparency, it’s the foundation that allows transparency to endure. @zama_fhe is proving this at the frontier of cryptography. By bringing Fully Homomorphic Encryption (FHE) to…

We’re officially partnering with @askalphaxiv to accelerate the future of open-source AI research. @askalphaxiv is building one of the most powerful open research platforms of our time, a space where anyone can search, explore, and understand the latest AI papers with clarity…

Be part of the first wave shaping how verified AI is built. Apply now for the chance to join the Perle Labs beta. Submissions close in 7 days 👇

The chain hums beneath the surface, a quiet rhythm of encrypted values moving, unseen, untouched, unstoppable. No eyes can watch them. No hands can alter them. Yet they think, they compute, they live. At the center of this silent storm stands @zama_fhe, the brain of the…

Being part of such a powerful community is an incredible feeling! In the biggest community movement in @SentientAGI history, this legendary raid, I was there too, and it’s truly an honor to be featured in the video. This is not just a video… It’s a memory of how we believed…

AI’s future won’t be controlled by a few, it will be powered by millions. @fortytwonetwork is turning everyday machines into a coordinated intelligence layer for the world. The Swarm Inference Protocol unlocks unused compute and transforms it into collective reasoning that can…

Blockchains proved that code can be law. But we learned something else along the way: Transparency without privacy can limit true freedom. This is where @zama_fhe changes the game. Fully Homomorphic Encryption (FHE) transforms blockchains into confidential computation networks,…

Sentient just achieved a major milestone at @NeurIPSConf, 4 research papers accepted across main tracks, benchmarks, competitions and security. This isn’t just a win, it’s a public validation of Sentient’s long-term vision for open, verifiable AGI. Each paper represents a…