Joe El Khoury·Jul 22Why AI Agents Forget — And Why That’s an Engineering ProblemMemory is not a model capability. It’s an infrastructure responsibility. And the teams that figure this out late don’t just fall behind —…A response icon2A response icon2
Joe El Khoury·Mar 22Why LLMs Hallucinate Rather Than Say “I Don’t Know”What information theory reveals about the fundamental limits of language models,and why retrieval-augmented architectures are not a…A response icon2A response icon2
Joe El Khoury·Mar 8How AI Agents and Ontologies Are Reshaping BankingIn 2025, Bank of America’s virtual assistant Erica quietly crossed 3 billion customer interactions — handling the equivalent workload of…
InGoPenAIbyJoe El Khoury·Aug 6, 2025The Missing Mug That Could Save AI: How World Models and RL Learn to Recognize Their Own…A mathematical approach that helps machines recognize when to ask for help
Joe El Khoury·Jul 12, 2025Dynamic Chunking (H-Net) : A NewApproach to Tokenizer-Free AI Text ProcessingThis article is written for AI researchers, Data scientists, machine learning engineers, NLP practitioners, and technology enthusiasts…
Joe El Khoury·Jun 15, 2025The Autonomy Paradox: Why Increasingly Capable AI Systems Require Stronger Human-Centered…This article explores a critical challenge in AI development: as artificial intelligence gains more independence, it creates proportionally…
Joe El Khoury·May 27, 2025Microsoft’s NLWeb: The Future of Conversational Websites?How a new open-source project could transform the web from static pages to intelligent conversationsA response icon1A response icon1
Joe El Khoury·Apr 3, 2025Why General-Purpose LLMs Won’t Modernize Your Codebase — And What WillAs organizations race to modernize legacy codebases, many are turning to Large Language Models (LLMs) as a potential silver bullet — only…A response icon2A response icon2
Joe El Khoury·Mar 7, 2025Masked Diffusion Models for Language- Based on LLaDA PaperIntroductionA response icon1A response icon1