Yan Hu

🫱 Introducing 𝐍𝐞𝐮𝐫𝐚𝐥 𝐂𝐨𝐦𝐩𝐮𝐭𝐞𝐫s: 𝐰𝐡𝐚𝐭 𝐢𝐟 𝐀𝐈 𝐝𝐨𝐞𝐬 𝐧𝐨𝐭 𝐣𝐮𝐬𝐭 𝐮𝐬𝐞 𝐜𝐨𝐦𝐩𝐮𝐭𝐞𝐫𝐬 𝐛𝐞𝐭𝐭𝐞𝐫, 𝐛𝐮𝐭 𝐛𝐞𝐠𝐢𝐧𝐬 𝐭𝐨 𝐛𝐞𝐜𝐨𝐦𝐞 𝐭𝐡𝐞 𝐫𝐮𝐧𝐧𝐢𝐧𝐠 𝐜𝐨𝐦𝐩𝐮𝐭𝐞𝐫 𝐢𝐭𝐬𝐞𝐥𝐟? Beyond today's conventional computers, agents, and world models, Neural Computers (NCs) are new frontiers where computation, memory, and I/O move into a learned runtime state. We ask: whether parts of runtime can move inward into the learning system itself. This is our first step toward the Completely Neural Computer (CNC): a general-purpose neural computer with stable execution, explicit reprogramming, and durable capability reuse. Work done with Mingchen Zhuge (@MingchenZhuge), Changsheng Zhao, Haozhe Liu (@HaoZhe65347 ), Zijian Zhou (@ZijianZhou524 ), Shuming Liu (@shuming96 ), Wenyi Wang (@Wenyi_AI_Wang ), Ernie Chang (@erniecyc ), Gael Le Lan, Junjie Fei, Wenxuan Zhang, Zhipeng Cai (@cai_zhipeng ), Zechun Liu (@zechunliu ), Yunyang Xiong (@YoungXiong1 ), Yining Yang, Yuandong Tian (@tydsh ), Yangyang Shi, Vikas Chandra (@vikasc), Juergen Schmidhuber (@SchmidhuberAI)

👏 ICLR 2026 **ORAL** - Thrilled to announce that our "Huxley-Gödel Machine" got accepted to ICLR 2026 as an ORAL presentation! @Wenyi_AI_Wang, @PiotrPiekosAI, @nbl_ai, Firas Laakom, @Beastlyprime, @MatOstasze, @MingchenZhuge, @SchmidhuberAI

While I was working on vision–language pretraining (VLP) in 2020—one of only4 VLP papers at CVPR 2021 (arxiv.org/pdf/2103.16110)—part of my thinking was that foundation models would refresh world models (arxiv.org/abs/1803.10122) proposed by @SchmidhuberAI and @hardmaru. When I was working on coding agents in 2023 (MetaGPT, among the first two LLM-based coding agents: arxiv.org/pdf/2308.00352), one of my core beliefs was that a practical version of @SchmidhuberAI’s Gödel Machine (recursive self-improvement) would be revitalized soon, as coding agents were rapidly becoming more powerful. In that work, we wrote the following in the outlook: “Generally speaking, a system should learn from experience in the real world, and meta-learn better learning algorithms from experiences of learning, and meta-meta-learn better meta-learning algorithms from experiences of meta-learning, etc., without any limitations except those of computability and physics.” We also worked on GPTSwarm in 2023 (gptswarm.org), which explored this direction to some extent. @DavidSilver and @RichardSSutton made learning from experience extremely popular in early 2025. In fact, as early as 2022, @SchmidhuberAI proposed the economy of minds (a multi-agent economy) in arxiv.org/abs/2305.17066. Recently, I’ve been developing new ideas looking toward 2027 and will release a paper soon. However, as I’m about to graduate, I may—or may have to—transition from (1)/(2) to (3). For me, all of (1), (2), and (3) are interesting 🙂 (1) and (2) need personal vision (3) needs vision from leadership

After a long journey of revisions and resubmissions, our paper “TwinMarket: A Scalable Behavioral and Social Simulation for Financial Markets” has finally been accepted to NeurIPS 2025. I’ll be at #neurips2025 in San Diego from Dec 1–8. Our poster session is on the morning of Dec 5. I’d be happy to discuss anything related to social agents, social/market simulations, and financial AI. We are also continuing to explore this direction and expect to share follow-up work in the future. Link: arxiv.org/pdf/2502.01506 Code: github.com/freedomintelli… Website: freedomintelligence.github.io/TwinMarket/

Thrilled to share that our paper TwinMarket received the 🏆 Best Paper Award at Financial AI Workshop @ ICLR 2025! Huge thanks to my amazing co-authors, my advisor Dr. Hu @sthuyan , and Prof. Wang @wabyking . Special thanks to the generous support from @RBCBorealis! #ICLR2025

[HAIE 2025-33] 헬스케어 AI 윤리 뉴스레터 blog.aiethics.dev/haie-2025-33/ 가장 흥미로운 논문으로 arXiv에 공개된 PrinciplismQA 벤치마크를 꼽고 싶습니다. 논문 저자는 LLM의 의료윤리적 판단을 시험하기 위해 "PrinciplismQA" 벤치마크 데이터를 개발했어요.

𝗧𝗼𝗽 𝘀𝗲𝗰𝗿𝗲𝘁 𝗔𝗴𝗲𝗻𝘁-𝗮𝘀-𝗮-𝗝𝘂𝗱𝗴𝗲 can be a great open-source #DeepWiki by just adding 2 code files swap github → openwiki on any repo URL 🫱 github.com/metauto-ai/age…

🚀 Excited to share TwinMarket! A scalable multi-agent framework using LLMs to simulate financial markets & emergent behavior 🎯📈 #Agent #LLM #Social 📄 Read more: huggingface.co/papers/2502.01…

We have a workshop at @iclr_conf ! 👏 👥 💡Our team members are taking part in organizing the 'Advances in Financial #AI workshop' at #ICLR2025. 🗓️ April 28, 2025 ⏰ 9:00am - 6:00pm SGT 📍 #Singapore EXPO, Hall 4 #2 Want to learn more? Visit ➡️lnkd.in/gh3tF6Dj 👋Swipe to meet the workshop organizers and speakers.

What are the best AI tools for research? Nature’s guide nature.com/articles/d4158… I discuss a little bit on deployment-efficient DeepSeek and o1/R1 -like reasoning for complex diagnosis.

Excited to share our work: Alignment at Pre-training! Towards Native Alignment for Arabic LLMs 🎉 📄 Published at NeurIPS ‘24 📍 NeurIPS East Exhibit Hall A-C #1803, Vancouver 🗓️ Fri 13 Dec 16:30 - 19:30 arxiv: arxiv.org/abs/2412.03253 poster: neurips.cc/virtual/2024/p…