The writing
Arguments, architectures, and practical questions about AI in organizations. Explore the writing by topic, or search for a specific question.
83 articles
Latent Reasoning in AI is transforming the way models process information by shifting from token-based reasoning to internal iterative computation. Unlike Chain-of-Thought (CoT) models, which verbalize every step, latent reasoning…
Discover how SmolLM2, a compact 1.7-billion parameter model developed by Hugging Face, redefines efficiency in language modeling. Unlike traditional large-scale models, SmolLM2 utilizes a data-centric training approach and multi-stage optimization…
Qwen2.5-1M: The First Open-Source AI Model with a 1 Million Token Context Window Qwen2.5-1M is a groundbreaking open-source AI model designed to process ultra-long documents with up to 1 million…
Retrieval-Augmented Generation (RAG) enhances AI by incorporating external knowledge, but optimizing its modules independently leads to inefficiencies. MMOA-RAG (Multi-Module Optimization Algorithm for RAG) solves this by using Multi-Agent Reinforcement Learning…
DeepSeek-R1 sets a new standard in artificial intelligence by leveraging a cutting-edge reinforcement learning (RL)-centric approach to enhance reasoning capabilities. Unlike traditional supervised fine-tuning methods, DeepSeek-R1 uses RL to autonomously…
Discover MiniMax-01, a groundbreaking AI model designed to overcome the limitations of traditional Large Language Models (LLMs) like GPT-4 and Claude-3.5. While current models handle up to 256K tokens, MiniMax-01…
Titans is a revolutionary neural architecture designed to overcome the limitations of traditional models like Transformers and recurrent networks. With its hybrid memory system integrating short-term, long-term, and persistent memory…
The Large Concept Model (LCM) introduces a groundbreaking approach to Natural Language Processing (NLP), transforming how machines understand and generate language. Unlike traditional token-based models, LCM focuses on concept-level understanding,…
AI hardware choices depend on the model, memory, interconnect, software, and operating conditions—not a chip label alone. This guide compares GPUs and specialized accelerators with edge, neuromorphic, photonic, and quantum…
Natural Language Processing (NLP) has long relied on tokenization as a foundational step to process and interpret human language. However, tokenization introduces limitations, including inefficiencies in handling noisy data, biases…
Artificial Intelligence (AI) has transformed how we interact with information, with Question Answering (QA) systems powered by Large Language Models (LLMs) becoming integral to decision-making across industries. However, challenges like…
Test Time Compute (TTC) represents a transformative shift in how AI systems process information, moving beyond traditional static inference to enable real-time adaptive reasoning. OpenAI's groundbreaking o1 model showcases this…