Published writing
A selection of explainers, articles, and newsletter pieces across AI, machine learning, and developer tools.
A comprehensive guide to GPT-4o Image Generation with 12 detailed prompt examples to help readers create the exact image they envision
Read on Learn PromptingTechnical analysis of ChatGPT's memory feature and its vulnerabilities to prompt injection attacks based on a recent real-world attack scenario.
Read on Learn PromptingEdited a practical guide to the key parameters that control LLM behavior: temperature, top-p, max tokens, frequency and presence penalties, and stop sequences, explaining how each one affects randomness, coherence, length, repetition, and novelty, with concrete examples using OpenAI's ChatGPT and API so you can reliably steer outputs from deterministic to creative.
Read on Learn PromptingCo-authored a guide comparing prompt engineering and fine-tuning as effective techniques for adapting LLMs to specific tasks.
Read on Learn PromptingInvestigation of a new “fun-tuning” attack on Google's Gemini models, showing how researchers used loss signals from the fine-tuning API to systematically bypass safety rules with increasingly effective prompt-injection attacks.
Read on Learn PromptingOverview of AI red teaming, its history and core concepts, how it differs from traditional security testing, the key components of a red team exercise, common vulnerabilities uncovered in practice, and the unique challenges faced by practitioners testing modern AI systems.
Read on Learn PromptingTechnical deep dive on JEPA (Joint-Embedding Predictive Architecture), a framework for self-supervised learning and world models proposed by Yann LeCun. Received public endorsement from Yann LeCun, its creator and one of the godfathers of AI, who called it an “Excellent blog post” on both LinkedIn and Twitter/X.
Read on Turing PostGuide to Graph RAG, explaining how graph-structured retrieval upgrades classic RAG by addressing its key limitations and leveraging knowledge graphs for better reasoning over complex information.
Read on Turing PostExplainer on the LongRAG framework, outlining the limitations of standard long-context RAG and introducing LongRAG's architecture and intuition. How it restructures retrieval to better handle long documents—along with its key advantages and curated resources for further study.
Read on Turing PostIntroduction to Mixture-of-Experts (MoE) models, tracing their history and original architecture, explaining why modular “experts” with conditional computation can outperform a single dense network, and showing how MoEs combined with Transformers underpin modern scalable systems like Mistral, DBRX, Jamba, Grok-1, and Arctic.
Read on Turing PostA piece on the elusive AI reading list Ilya Sutskever reportedly gave John Carmack, tracing the origin of this “90% of what matters today” claim and how the unseen ~40-paper list became a legend in the ML community.
Read on Turing PostExplainer on hallucinations in foundation models - what they are, why they occur, how to detect them, when they're harmful or occasionally useful. Practical strategies to reduce or mitigate them, and a curated set of tools, libraries, and datasets for working with hallucinations in practice.
Read on Turing PostGuide to Low-Rank Adaptation (LoRA) in the context of large language models, explaining when full fine-tuning is truly necessary, how LoRA compares to other adaptation techniques, the core intuition behind it, how it works under the hood, and why it offers a more efficient way to specialize billion-parameter LLMs.
Read on Turing PostAuthored the technical sections on Transformer-based and Diffusion-based models and conducted research on generative model architectures to trace the origins of diffusion models; other generative models aren't covered due to space constraints.
Read on Turing PostSurvey of the Chain-of-Thought Prompting lineage, tracing how the original NeurIPS 2022 Google Brain paper sparked a wave of follow-up work and “chain”-style methods, from Self-Consistency, Zero-Shot CoT, Auto-CoT, and Program-of-Thoughts to Multimodal CoT, Tree-of-Thoughts, Graph-of-Thoughts, Algorithm-of-Thoughts, and Skeleton-of-Thought, and mapping their core ideas, innovations, and relationships in a single, up-to-date reference.
Read on Turing PostPrimer on Retrieval-Augmented Generation (RAG), explaining what it is, which core LLM limitations it addresses, how its architecture works, and why it has become so popular.
Read on Turing PostGathered expert commentary on handling missing values in machine learning datasets, featuring insights from three industry experts: Ryan Kearns (Founding Data Scientist, Monte Carlo), David Berenstein (Developer Advocate, Argilla), and Abhishek Pawar (Senior Data Scientist, Precisely).
Read on Turing PostGathered expert commentary on implementing open-source LLMs in production, featuring insights from five industry experts: Edward Beeching (Hugging Face research scientist and co-creator of Open LLMs leaderboard), Rajiv Shah (Hugging Face ML engineer), Aniket Maurya (Lightning AI developer advocate), Lianmin Zheng (UC Berkeley PhD student and Vicuna contributor), and Devis Lucato (Microsoft principal architect, Semantic Kernel).
Read on Turing PostDeep dive into three leading open-source LLMs according to Hugging Face's Open LLM Leaderboard: LLaMA, LLaMa 2 and Falcon, and their top fine-tuned variants
Read on Turing PostHistorical deep dive into the AI and ML breakthroughs of the 1990s to mid-2000s, showing how foundational ideas from that era, developed under severe computational constraints, paved the way for today's large language models.
Read on Turing PostExploration of the origins of anthropomorphism in AI, examining why humans instinctively attribute human qualities to non-human systems, how fictional narratives and media portrayals of robots and AI shape public perception, and the psychological implications of these tendencies for human–AI relationships and trust.
Read on Turing PostFrom autonomous AI and multimodal capabilities to smaller language models, accelerated product development, and chatbots with near-infinite memory.
Read on Learn PromptingMake the most out of your chatbot or language model by learning prompt engineering with these online courses.
Read on Learn PromptingHighlights from Dario Amodei's 5-hour Lex Fridman Podcast on AI's challenges, safety, and scaling.
Read on Learn PromptingLearn how to write effective prompts to get precise and useful responses from AI models with these 10 expert tips.
Read on Learn PromptingThe Search Engine You Didn’t Know Could Do So Much
Read on Learn PromptingHow to use Apple Intelligence? Simplify tasks, summarize messages, rewrite text, and create custom emojis. Smarter, faster, and more fun on Apple devices!
Read on Learn PromptingDiscover Google Gemini, an AI chatbot that uses multimodal capabilities and Google Search to simplify tasks like handling text, images, and audio.
Read on Learn PromptingDiscover how Suno AI makes music creation simple and fun. From custom soundtracks to personalized gifts, unleash your creativity with ease.
Read on Learn PromptingIn this week's newsletter, we're diving into the latest updates from OpenAI, Anthropic, Google, and more. Plus, we've got a special announcement about our upcoming AI Red Teaming and AI Security Masterclass.
Read on Learn PromptingDiscover how researchers are creating simulation agents that can replicate human attitudes and behaviors with 85% accuracy, and learn about the latest developments in AI agents from Microsoft, Google, and more.
Read on Learn PromptingPlus, updates from Perplexity, Replit, ElevenLabs, and Google, must-have tools for developers, and 2 papers on the real impact of prompting—is Prompting dead or thriving?
Read on Learn PromptingJoin HackAPrompt 2.0, the largest AI red-teaming competition ever! Test vulnerabilities in AI systems across 5 tracks with $500K in prizes.
Read on Learn PromptingPlus, Learn Prompting x US Congress, HackAPrompt 2.0 Updates, & AI Red Teaming Course
Read on Learn PromptingDiscover how Apple integrates Responsible AI principles into the development of Apple Foundation Models (AFM), the driving force behind Apple Intelligence.
Read on Learn PromptingDiscover the top free AI image generators of 2025! Learn how these tools can help you create stunning visuals effortlessly, no artistic skills required.
Read on Learn PromptingPlus, new updates from Nvidia, Microsoft, and Adobe!
Read on Learn PromptingGoogle has announced its largest acquisition to date, purchasing cloud security company Wiz for $32 billion in a strategic move to strengthen its position in the cloud computing market.
Read on Learn PromptingA comprehensive study by researchers at Stanford, the University of Washington, Emory University, and others reveals how AI tools are transforming consumer complaints, corporate communications, job postings, and even United Nations press releases.
Read on Learn PromptingPlus, the latest updates on generative AI tools from xAI, Perplexity, Adobe, OpenAI, and more.
Read on Learn PromptingExplore how Apple ensures AI privacy with on-device processing and Private Cloud Compute, designed to keep your data safe.
Read on Learn PromptingLearn how to harness ChatGPT's Canvas tool for collaborative editing, text refining, coding assistance, and streamlined project workflows.
Read on Learn PromptingYour Ultimate Guide to Resume Creation with AI
Read on Learn PromptingDiscover how to integrate real-time web data into ChatGPT for timely, sourced answers, streamlining research and eliminating tab hopping.
Read on Learn PromptingLearn to engage with ChatGPT through natural speech, share visuals, and run code interactively, enhancing conversations with richer, dynamic content.
Read on Learn PromptingOrganize chats, files, and instructions into structured folders, creating tailored workspaces that streamline complex projects and ensure context.
Read on Learn PromptingPlus, a Brief History of AI at Apple
Read on Learn PromptingLasso Security's investigation reveals Microsoft's AI assistant can access private GitHub repositories through cached data, raising major security concerns
Read on Learn PromptingResearchers discovered new ways to trick AI assistants in blockchain systems. Learn how these attacks work and why they matter for the future of AI security.
Read on Learn PromptingA comprehensive guide to OpenAI's new Agents SDK, exploring its features, benefits, and practical applications for building AI-powered applications.
Read on Learn PromptingOpenAI's GPT-4o is turning heads with image generation capabilities that come remarkably close to professional quality. The model creates visuals comparable to professional services but does it in minutes instead of hours and at a fraction of the cost. This is both exciting and a bit concerning for the creative industry.
Read on Learn PromptingA detailed analysis of how ChatGPT Operator can be exploited through prompt injection, including real-world attack scenarios and defense mechanisms.
Read on Learn PromptingSakana AI's AI Scientist-v2 recently achieved a major milestone by generating its first fully AI-created scientific paper that passed the peer-review process at an ICLR workshop.
Read on Learn PromptingSesame's Conversational Speech Model (CSM) is a breakthrough in AI speech generation, enabling more natural and engaging conversations with AI.
Read on Learn PromptingSesame AI has open-sourced their Conversational Speech Model (CSM), making advanced speech synthesis technology available to developers and researchers.
Read on Learn PromptingStability AI releases Stable Virtual Camera, a breakthrough multi-view diffusion model that converts still images into 3D videos with sophisticated camera control capabilities.
Read on Learn PromptingTavus introduces a new Conversational Video Interface (CVI) that enables AI agents to perceive, listen, understand, and engage in dynamic, human-like conversations.
Read on Learn PromptingSpanning AI research, industry trends, politics, and safety.
Read on Learn PromptingFrom Google's Veo to OpenAI's Sora, the multimodal AI space is heating up. Plus, explore the future of AI and learn GenAI with 15 hands-on courses!
Read on Learn PromptingExplore 15 courses for learning how to master ChatGPT for everyday tasks, both personal and professional.
Read on Learn PromptingDiscover the 20 best online courses on generative AI, suitable for all experience levels and budgets.
Read on Learn PromptingExplore the top 29 online courses to master Large Language Models (LLMs), including free and paid options, covering everything from basics to advanced techniques.
Read on Learn PromptingLarge Language Diffusion with mAsking (LLaDA) represents a transformative shift in how language generation can be approached, by leveraging diffusion techniques originally popularized in image and audio domains.
Read on Learn PromptingExplore DeepSeek-R1, the open-source AI model matching proprietary systems through innovative reinforcement learning and advanced reasoning capabilities.
Read on Learn PromptingGoogle Agentspace unites enterprise data, search, and AI agents, offering contextual insights, streamlined workflows, and potential efficiency gains.
Read on Learn PromptingManus is a new kind of general AI agent designed to go beyond chat-based interactions, delivering real-world results without continuous supervision.
Read on Learn PromptingLearn how prompt engineering works, why it matters, and how to get started. Understand the importance of phrasing your instructions, and explore techniques like few-shot prompting, chain-of-thought, and role prompting.
Read on Learn PromptingCompete in HackAPrompt 2.0, the world's largest AI Red-Teaming competition!
Read on Learn PromptingExplore vibe coding, the emerging trend where developers 'give in to the vibes' and let AI handle most coding tasks. Learn how this approach is transforming software development and what it means for the future of programming.
Read on Learn PromptingIn this week's newsletter, we're diving into the latest updates from OpenAI, Anthropic, Google, and more. Plus, we've got a special announcement about our upcoming AI Red Teaming and AI Security Masterclass.
Read on Learn PromptingCo-authored profile of xAI, the AI company founded by Elon Musk.
Read on Turing PostCo-authored comprehensive profile of Cerebras, the AI chip company revolutionizing large-scale AI training with their wafer-scale engines and innovative hardware architecture.
Read on Turing PostAuthored technical guide to Mamba, the state-space model architecture that challenges Transformer dominance with linear scaling.
Read on Turing PostCo-authored deep dive into Moonshot AI, the Chinese unicorn pioneering breakthrough long-context AI technology with their Kimi chatbot and 200K+ token context windows.
Read on Turing PostCo-authored profile of Zhipu AI, one of China's leading AI companies behind the ChatGLM series and innovative multimodal AI solutions.
Read on Turing PostCo-authored analysis of Snowflake's evolution from cloud data warehouse to AI platform, exploring their machine learning capabilities and data cloud ecosystem.
Read on Turing PostCo-authored historical exploration of computer vision breakthroughs in the 1980s, covering key algorithms and foundational research.
Read on Turing PostAuthored historical journey through early computer vision development from 1950-1970s, covering foundational concepts and pioneers.
Read on Turing PostCo-authored comprehensive history of computer vision evolution and its crucial role in the journey toward artificial general intelligence.
Read on Turing PostAuthored guide to multimodal AI models, exploring how systems integrate text, vision, and audio for comprehensive understanding.
Read on Turing PostCo-authored profile of CoreWeave, the GPU cloud infrastructure company powering AI workloads with specialized hardware for machine learning and rendering.
Read on Turing PostAuthored comprehensive guide to data privacy challenges in large language model systems, covering compliance, security, and best practices.
Read on Turing PostAuthored comprehensive profile of Scale AI, the $7.3B data infrastructure company powering AI training for autonomous vehicles, defense, and enterprise applications.
Read on Turing PostCo-authored profile of Databricks, exploring their lakehouse architecture, unified analytics platform, and leadership in enterprise AI and machine learning workflows.
Read on Turing PostAuthored curated list of the best free courses for learning large language models, from fundamentals to advanced applications.
Read on Turing PostCo-authored profile of AI21 Labs, the Israeli startup behind Jurassic LLMs and innovative AI writing tools, exploring their unique approach to language models.
Read on Turing PostAuthored a deep dive into Mamba's architecture and how it addresses the computational limitations of transformer models.
Read on Turing PostAuthored curated list of 8 essential open-source tools for implementing RAG systems, from vector databases to orchestration frameworks.
Read on Turing PostAuthored analysis of the trade-offs between large and small language models, exploring efficiency, performance, and deployment considerations.
Read on Turing PostAuthored comparative analysis of open-source versus closed-source AI models, examining benefits, risks, and strategic implications.
Read on Turing PostCo-authored a piece about Chain-of-Verification, Chain of Density, and Self-Refine.
Read on Turing PostCo-authored foundational guide to foundation models, explaining their architecture, training, and role as building blocks of modern AI.
Read on Turing PostCo-authored profile of Lightricks, the creative technology company behind popular apps like Facetune, exploring their AI-powered photo and video editing innovations.
Read on Turing PostCo-authored analysis of Cohere, the enterprise-focused AI company providing language models and NLP solutions for business applications and retrieval systems.
Read on Turing PostAuthored a piece about the history of LLMs starting from pre-trained models to GPT-4.
Read on Turing PostAuthored a piece exploring the history of AI winters, periods of reduced funding and interest in AI research.
Read on Turing PostCo-authored a piece exploring AI and computer science research from the late 1950s to around 1970.
Read on Turing PostCo-authored history exploring The Era of Mechanical Translation and How It Crashed (History of LLMs #1) and their impact in the AI landscape.
Read on Turing Post