behind the scenes - the making of natives font https://t.co/OSUKvsBmzH

Integrating LLMs with knowledge bases. Important read for AI practitioners LLMs generate impressive text but struggle with hallucinations, outdated knowledge, and reasoning over structured data. The default response has been scaling up (e.g., more parameters, more compute, more cost). But bigger models don't solve the fundamental problem: LLMs lack reliable access to external, verifiable knowledge. This new survey examines how RAG, Knowledge Graphs, and hybrid approaches address these limitations. The key insight: integration happens at three levels: - Level 1 focuses on retrieval, getting the right information into the model. - Level 2 addresses reasoning, synthesizing retrieved knowledge for complex tasks. - Level 3 handles optimization, adapting systems for domain-specific needs. KAG showed 19.1% exact match improvement over basic RAG on HotpotQA. Think-on-Graph achieved significant accuracy gains over Chain-of-Thought on complex QA. The practical applications span finance, medicine, and code generation. FinAgent combines RAG with reinforcement learning for trading decisions. UMLS integration improves diagnostic accuracy in medical AI. Codex leverages retrieval to enhance code generation quality. Knowledge drift requires continuous updates, domain-specific representations don't always align with LLM embeddings, and standardized evaluation benchmarks are still lacking. The path to reliable LLMs isn't just scale. It's thoughtful integration with structured knowledge that provides factual grounding and enables complex reasoning. Paper: https://t.co/vl8ZPf4ncA Learn to build RAG and AI agents in our academy: https://t.co/zQXQt0PMbG

🦸1Wrk’s SuperApp, built on #AWS, is helping businesses hire faster & mange talent more intelligently. 💡Discover how AWS is providing a “backbone” for #generativeAI innovation & helping the #startup scale: https://t.co/hxJI58TMJ9 https://t.co/8Ij87Ipbpb

📣 @DeepgramAI announces integration with key #AWS solutions at #AWSreInvent. 🛠 Accessing advanced voice #AI models & capabilities within their existing #AWS environments will enable customers to build, deploy & scale their applications faster & more securely. 🔗 https://t.co/FiXYywbF0s

🙌 #awsreinvent is back & Day 1 is done! https://t.co/PZYZLKBKHV 🦾 We met the next generation of leaders, talked about the latest trends in #AI & celebrated opening night in style with the global cloud community. What a day for #strartups! https://t.co/XMhbIyRxzI

📣 Applications for the 2026 Physical AI Fellowship are now open! ✅ https://t.co/Wvje7yrqsy 🤖 At #awsreinvent, @MassRobotics & @nvidia announce the second edition of the 8-week program designed to help robotics & physical #AI startups from around the world scale faster & smarter.

📣 @twelve_labs launches its most powerful video understanding model at #awsreinvent. https://t.co/jEHEFPJYQe 🎞 Marengo 3.0 “shatters the limits of what’s possible” for developers & enterprise, enabling them to search, navigate & understand video content at scale. https://t.co/ogx3CtxAJa

🤝 Networking, exploring the future with Matt Garman & celebrating success: what a fantastic day at #awsreinvent! ⚡️The energy was electrice, with Day 2 bringing innovation & insights. 🙌 Keep it up, cloud community! https://t.co/KzKSpjirk6 https://t.co/NaIVNBZBXA

🌱 At #awsreinvent, @BrainBoxAI announces the expansion of its AI pilot for sustainability, with #AWS. 🔋 The project will scale across the US, improving energy efficiency & decarbonization at #Amazon fulfillment sites. https://t.co/5cBry2jJ4J https://t.co/ZcZNEZeytD

🎧🔈 Using AI & machine learning to "unmix” audio recordings, @AudioShake makes audio workflows clear & easy to control. 👉 https://t.co/tEDUnJIANA Backed by #AWS infrastructure, AudioShake is able to process large workloads while staying flexible & fast. https://t.co/xG4ymlPPQ6

📣🔏 @Get_Writer launches new agent supervision & orchestration tools with #AWS. At #awsreinvent, the startup announced its integration with #AmazonBedrock, enabling enterprises to easily build secure agents & get wider access to #AI models. 👉 https://t.co/cdLnaXoHm0 https://t.co/eKx7VYsl9p

🌏 Day 3 of #awsreinvent was buzzing with AI insights & our VP of Startups Jason Bennett shares the highlights. 💡 From a showcase of bold ideas to a keynote from VP of Agentic #AI @SwamiSivasubram —we saw the future of transformative tech. https://t.co/SYabp5Nvmd https://t.co/FdFScnBcrw

🤖 @agilityrobotics is redefining the warehouse workforce with its humanoid robot, Digit. Discover how the startup continuously optimizes & scales its AI model training through #AWS Cloud, driving both productivity & wellness. 👉 https://t.co/jaQ7JmSLZf https://t.co/jFQFuCIbp4

🙌#awsreinvent: what a week! 🌍Future leaders pitched bold ideas, #startups built global connections, investors shared insights & pioneers unpacked the next era of #AI. 👋A huge thank you to the amazing cloud community—see you all next year! https://t.co/uepoI953f6 https://t.co/x4z0RCnDvL

@FPLGOAT7 Ranked 4,000 GW1, currently 49k. Any suggestions? https://t.co/EmCYm2yftH

Great overview of Nvidia open source goodies for researchers @NaderLikeLadder https://t.co/VMjG3DhzJD

Great overview of Nvidia open source goodies for researchers @NaderLikeLadder https://t.co/VMjG3DhzJD

I'm at NeurIPS! Giving a talk today on state of OSS & why NVIDIA spends millions of GPU hours training models just to give it all away: datasets, recipes, weights, architecture. 545pm @ Exhibit Hall A,B https://t.co/G3SBf9iHr8

Me & my 2 acquirers 💚💍 https://t.co/UXurHHEz9D

+2 more

If you're building or experimenting with GenAI on Arm devices, there’s a great learning session coming up next week! Join Gian Marco Iodice, Principal Software Engineer at @Arm, and Digant Desai, Software Engineer at @Meta, as they unpack the latest updates to ExecuTorch, including SME2 integration and the OpenSmall model. They’ll walk you through how these additions make it easier to run GenAI workloads consistently across edge, mobile, and cloud environments. Expect practical examples, code walkthroughs, and open discussion with the engineers who work on this every day. 📎 Dec 11, 9 am (next Thursday): https://t.co/t3SHH1UHy5 #PyTorchFoundation #OpenSourceAI #ExecuTorch

Need to accelerate Large-Scale Mixture of Experts Training? With @nvidia NeMo Automodel, an open source library within NVIDIA NeMo framework, developers can now train large-scale MoE models directly in PyTorch using the same familiar tools they already know. Learn how to make MoE training of massive models simple and efficient by reading NVIDIA’s developer blog: 📎 https://t.co/xGnx95jY8y #PyTorch #MoE #OpenSourceAI

I just read the DGM paper by @jennyzhangzt et al. This is one of those papers where I can't wait to actually work through the whole appendix! After discovering CTMs and the AI-Scientist I'm struggling to get over all the cool shit that seems to be coming from @SakanaAILabs! https://t.co/jH3Rl16Mro

【経営企画メンバー募集！】 Corporate Development https://t.co/Iik4DnIYAf Sakana AIは、研究開発と事業開発の両輪で成長をとげ、日本発でありながら既にグローバルAI企業や投資家からも注目される存在となっています。この度、当社の成長をさらに進化・スケールさせるため、「AI技術 x 産業変革 x 資本戦略」の交差点で戦略を一緒に描いてくれる 経営企画メンバーを募集します！ CEO/COOと密に連携し、資本政策、M&A、戦略的パートナーシップなどを通じて、当社の次の成長を一緒に創り上げてくれる方をお待ちしています。

We are expanding Sakana AI’s corporate development team. If you have worked in Corp Dev, IBD at a global investment bank in Japan, or PE in Japan, please have a look. Business Role (Corporate Development): https://t.co/2VbQQ7FCzw

Cool workshop at #NeurIPS2025 AI for Non-Human Animal Communication https://t.co/9Q7UW0rJ85

LAW 2025: Bridging Language, Agent, and World Models Looks interesting, seems @fchollet @danijarh and others are speakers at this workshop at #NeurIPS2025 https://t.co/raeg3DHDLI

I love seeing all the amazing things our Google Brain Residents and AI Residency Program cohorts have gone on to do! @hyhieu226's post spurred me to dig up a blog post and video from that era. Blog: "Google Brain Residency Program - 7 months in and looking ahead" https://t.co/KvXa5xICp5 Video: https://t.co/URT43V20Mn

So much memory from the first class of Google Brain Residency. It was a good year: we were all young and free. We had no idea what makes money and what does not. No bs politics. No promotion tricks. We just did what we liked. https://t.co/SDXUjSNJcE

Year end review: 2025 was dominated by LLM research! https://t.co/TsoX4KPN73

@pmddomingos Spotted you at the Music Workshop 🎹 https://t.co/cqEZleKXdH

Anthropic research shows the insider threat of agentic misalignment https://t.co/HUT0LE55KP by @bendee983 via @bdtechtalks #GenerativeAI #AI #MachineLearning Cc @jblefevre60 @SpirosMargaris @HaroldSinnott @Ym78200 @PawlowskiMario https://t.co/skdklkWgbV

Anthropic research shows the insider threat of agentic misalignment https://t.co/HUT0LE55KP by @bendee983 via @bdtechtalks #GenerativeAI #AI #MachineLearning Cc @jblefevre60 @SpirosMargaris @HaroldSinnott @Ym78200 @PawlowskiMario https://t.co/skdklkWgbV