Intel
Page 24
Controlling Reasoning Effort in LLMs
An in-depth technical article explaining how large language models learn to support low-, medium-, and high-effort reasoning modes.
Why it mattersIf you tune reasoning effort in models like GPT-5, Qwen3, or gpt-oss, this deep-dive explains how those low/medium/high modes are actually trained and implemented under the hood via RLVR and think tokens.
articleSebastian Raschka, PhD
Neo Kim@systemdesignone
A comprehensive course for getting into Large Language Models (LLMs), featuring structured roadmaps and hands-on Google Colab notebooks. It walks learners through the full LLM stack from fundamentals to advanced topics. Ideal for developers and researchers looking to build practical skills in modern AI.
Why it mattersIf you want to go from zero to deploying production LLM applications, this course provides a structured three-part path — fundamentals, model science, and engineering.
tobi lutke@tobi
A personal essay by Shopify CEO Tobi Lütke about his apprenticeship in Germany after dropping out of school at 16 to work at a Siemens subsidiary. The piece reflects on how hands-on, basement-level hacker culture shaped his technical and entrepreneurial thinking.
Why it mattersRiver shows what happens when you design an AI coding agent around organizational transparency rather than individual productivity.
Satya Nadella@satyanadella
An article shared by Microsoft CEO Satya Nadella exploring how firms should protect their core intellectual property in the age of AI and intelligence. It references Nobel Prize-winning economist Kenneth Arrow's information paradox and examines its inverse implications for modern businesses.
Why it mattersIf you're building with AI at a firm that has proprietary workflows, customer data, or hard-won operational knowledge, every prompt you send is potentially training your competitor's future product.
Imade.@ImadeIyamu
Billion Dollar PDFs is a curated collection of exceptional writing and documents available online. It aggregates high-quality PDFs — likely including influential business, finance, and intellectual works — making them easy to discover and access in one place.
Why it mattersIf you've ever wanted to understand *why* markets moved the way they did, primary documents tell the story better than any retrospective analysis.
Grey@greyluox
An article by Grey (@greyluox) arguing that while AI has dramatically reduced the cost of producing software, it has not reduced the cost of building the wrong thing. The piece explores how teams are responding to cheaper implementation by making larger bets, potentially amplifying mistakes.
Why it mattersIf you're using AI to ship faster and finding yourself with large codebases that don't quite fit the problem, this article gives you a decision rule: match your implementation batch size to your uncertainty level.
Daniel Kornum@dkfromdk
An article arguing that the true winners of the AI era will be low-margin businesses — not high-tech companies — because AI dramatically reduces their operational costs and unlocks massive efficiency gains. It reframes the AI competitive landscape around cost structure rather than engineering headcount or data budgets.
Why it mattersIf your business runs on thin margins — distribution, field service, logistics.
Chamath Palihapitiya@chamath
An article by Chamath Palihapitiya exploring the concept of a 'software factory' — a modern approach to building software at scale. It draws from real-world experience at 8090, where the team reverse-engineered and replaced an 18-million-line COBOL/Assembly billing engine.
Why it mattersIf you're building or evaluating AI-assisted development pipelines, this piece gives you the historical and conceptual vocabulary to distinguish real industrial-grade systems from the current wave of hype.
Pierson Marks@piersonmarks
A personal software factory system built to automate software creation workflows with simplicity as a core principle. The creator shares how it works, what they learned, and how others can replicate it in an afternoon. It's a practical guide-plus-system for anyone looking to systematize AI-assisted software development.
Why it mattersIf you're trying to run AI agents on real software work without babysitting them constantly, the insight here is structural.
Neil Rahilly@neilrahilly
An article by Neil Rahilly about adopting AI tools across a company, framed through the lens of the 1968 study on software engineer productivity variance. It explores why AI adoption matters and how it can dramatically amplify individual and team output.
Why it mattersIf you're trying to get your whole company using AI — not just engineers — the biggest trap is building one agent per department.
The Pulse: What can we learn from Bun’s rapid Rust rewrite with AI?
Spending $165,000 to migrate Bun from Zig to Rust looks expensive until you note it compressed a one-to-two-year migration into 11 days.
Why it mattersMost AI-coding takes stay theoretical; this one crunches the actual numbers on a real migration — $165K and 11 days to move Bun from Zig to Rust, versus the 1-2 years it would've taken by hand.
articleGergely Orosz
Zara Zhang@zarazhangrui
An article by Zara Zhang exploring how to shed school-conditioned mental blocks around learning, using language learning as a case study. It challenges the 'bad at languages' self-label and offers a reframe for adult learners returning to skills they abandoned. Written for anyone who wants to learn something new but feels held back by past academic failure.
Why it mattersIf you've ever abandoned a learning project because the 'start from basics' approach killed your motivation, this reframe — start with a real problem, work backwards to the knowledge you need, use AI as a 24/7 tutor.
Vlad Feinberg@FeinbergVlad
A guide by Vlad Feinberg (Gemini Pretraining Area TL at Google DeepMind) on how to get hired at a top AI research lab like Anthropic, OpenAI, xAI, or DeepMind. It offers insider perspective from someone actively working at the frontier of AI research and development.
Why it mattersIf you're targeting a job at OpenAI, Anthropic, or Google DeepMind, most advice you'll find online comes from recruiters or people who interviewed there once.
ClaudeDevs@ClaudeDevs
An article by ClaudeDevs explaining the concept of 'designing loops' for coding agents, a emerging paradigm shift from traditional prompting. It helps developers building with Claude AI understand how to structure agent workflows for more reliable and effective code generation.
Why it mattersIf you're building agentic workflows with Claude Code, this is the official taxonomy straight from the team building the tool — not another Twitter thread interpretation.
Moish Peltz@mpeltz
A public-domain collection of generative AI policy templates designed to help organizations quickly create and adopt AI governance policies. Built by a digital assets and IP law firm, these templates provide legally-informed starting points for businesses navigating AI compliance. Free to use and adapt without copyright restrictions.
Why it mattersIf your law firm or small business needs an AI governance policy but can't justify a consultant engagement, this gives you battle-tested, lawyer-authored templates in public domain.
articleLilian Weng
Harness Engineering for Self-Improvement
Recursive self-improvement dates to I. J. Good in 1965 and Yudkowsky's 2008 framing of the feedback loop.
Why it mattersWeng reframes the AI coding conversation from prompting tricks to the harness itself.
Daniel Fried@dan_fried
A university course at Carnegie Mellon University focused on building AI agents end-to-end. Students gain hands-on experience constructing agentic harnesses, building evaluations, and training LLMs with reinforcement learning. The course balances theory and practice while covering modern frameworks and best practices.
Why it mattersIf you're building AI agents and want to go beyond tutorials, this CMU graduate course offers a structured, research-backed curriculum covering the full stack of agent design — tool use, memory, planning, and safety.
Thariq@trq212
An article by Thariq exploring how to work effectively with Claude's Fable system, drawing on the insight that 'the map is not the territory.' It guides readers on understanding the gap between their prompts/context and the actual work being done by AI. Useful for developers and creators trying to get more out of AI-assisted workflows.
Why it mattersIf your AI coding sessions keep drifting off-track, the problem usually isn't the model — it's unresolved unknowns you didn't know you had.
ellington@not_ellington
A comprehensive practitioner's reference book hosted on arXiv covering the full stack of building autonomous AI systems, from foundational principles to production deployment. It's positioned as a deep, substantive alternative to the shallow 'edu-slop' threads and quick-read content proliferating on social media.
Why it mattersA practitioner's reference on arXiv covering the full stack of agentic AI, from first principles to production deployment.
The twilight of the chatbots
Ethan Mollick on the chatbot receding as the main way people meet AI, as capability gains accelerate and the frontier stays jagged even while release cadence quickens.
Why it mattersMollick's argument isn't another 'prompt better' guide — it's a data-grounded case that the chatbot-as-interface era is ending, with capability gains accelerating even as the frontier stays jagged and unreliable in specific spots.
articleEthan Mollick
Latest open artifacts (#22): Zyphra, Cohere, and Poolside are expanding the breadth of the ecosystem
A roundup of recent open model releases showing the ecosystem diversifying well beyond the handful of mostly Chinese labs that dominated it a year ago.
Why it mattersMost open-model coverage fixates on the frontier race, but this tracks a quieter structural shift: the release landscape has broadened from a handful of mostly-Chinese labs to niche players worldwide like Zyphra, Cohere, and Poolside.
articleFlorian Brand
Lenny Rachitsky
OpenAI Codex lead on the new shape of product work | Andrew Ambrosino
podcast
Listen
Using Local Coding Agents
A tutorial for assembling a fully local coding agent from open-source tools and open-weight models, pairing an agent harness with a model served by a local inference runtime.
Why it mattersMost agent tutorials assume you're piping code to a frontier API; Raschka instead wires up the whole loop locally — open-weight models served through an inference runtime, driving a real coding agent harness.
articleSebastian Raschka, PhD
Lenny Rachitsky
What happens after coding is solved? | Fiona Fung (Manager of the Claude Code and Cowork Teams)
podcast
Listen
LLM Research Papers: The 2026 List
A curated, categorized reading list of notable LLM research papers published from January to May 2026, organized into topics like architecture, reasoning, reinforcement learning.
Why it mattersA single well-organized snapshot of the most notable early-2026 LLM papers grouped by topic (architecture, reasoning, RL, agents).
articleSebastian Raschka, PhD
Lenny Rachitsky
The AI paradox: More automation, more humans, more work | Dan Shipper
podcast
Listen
Recent Developments in LLM Architectures: KV Sharing, mHC, and Compressed Attention
An in-depth technical article analyzing recent open-weight LLM architecture innovations focused on long-context efficiency.
Why it mattersIf you're building or optimizing long-context LLM inference, this breaks down how the newest open-weight models actually shrink KV cache and attention costs.
articleSebastian Raschka, PhD
Lenny Rachitsky
Why cultivating agency matters more than cultivating skills in the AI era | Max Schoening (Head of Product, Notion)
podcast
Listen
Lenny Rachitsky
How Anthropic’s product team moves faster than anyone else | Cat Wu (Head of Product, Claude Code)
podcast
Listen
Understanding LLM Architectures Workflow
A learning-oriented walkthrough of how to reverse-engineer and understand new open-weight LLM architectures by inspecting Hugging Face config files and reference implementations rather than relying on sparse technical papers.
Why it mattersIf you want to genuinely understand how a new open-weight model works, this shows you how to decode Hugging Face config files and reference implementations directly — trusting working code over increasingly vague technical papers.
articleSebastian Raschka, PhD
Lenny Rachitsky
Head of Growth (Anthropic): “Claude is growing itself at this point” | Amol Avasare
podcast
Listen
Mini Coding Agent
A minimal but fully working coding agent implemented in pure Python, designed to demonstrate the six core building blocks of coding agents like Claude Code and Codex CLI.
Why it mattersIf you want to understand how tools like Claude Code and Codex CLI actually work rather than just use them, this breaks the agent harness into six concrete building blocks and backs each with a minimal, readable Python implementation you can study end-to-end.
articleSebastian Raschka, PhD
Lenny Rachitsky
An AI state of the union: We’ve passed the inflection point, dark factories are coming, and automation timelines | Simon Willison
podcast
Listen
A Visual Guide to Attention Variants in Modern LLMs
An educational article and visual gallery explaining the evolution of attention mechanisms in modern LLMs, from Multi-Head Attention (MHA) and Grouped-Query Attention (GQA) to Multi-Head Latent Attention (MLA) and sparse/hybrid architectures.
Why it mattersIf you want to understand how modern LLMs actually manage attention — and why models are shifting to GQA, MLA, and sparse/hybrid schemes.
articleSebastian Raschka, PhD
Categories of Inference-Time Scaling for Improved LLM Reasoning
An in-depth educational article that categorizes and explains inference-time scaling techniques for improving LLM reasoning.
Why it mattersIt gives a clear, practical taxonomy of inference-time scaling methods with code and experimental results.
articleSebastian Raschka, PhD
The State Of LLMs 2025
An in-depth year-in-review article covering the major developments in large language models throughout 2025, including DeepSeek R1, RLVR, GRPO, inference-time scaling.
Why it mattersIt distills a chaotic year of LLM research — DeepSeek's cost disruption, the rise of RLVR/GRPO for reasoning, and inference-time scaling.
articleSebastian Raschka, PhD
A Technical Tour of the DeepSeek Models from V3 to V3.2
An in-depth technical article walking through the architectural evolution of DeepSeek's open-weight LLMs from V3 to V3.2, explaining concepts like Multi-Head Latent Attention, RLVR reasoning training.
Why it mattersIf you're evaluating open-weight LLMs for agentic work, this walks through DeepSeek's architectural evolution from V3 to V3.2 — including the DeepSeek Sparse Attention mechanism and RLVR reasoning training.
articleSebastian Raschka, PhD
Beyond Standard LLMs
An in-depth educational article by Sebastian Raschka exploring alternative LLM architectures beyond standard autoregressive transformers, including linear attention hybrids, text diffusion models, code world models, and small recursive transformers.
Why it mattersIf you track LLM architecture trends, this breaks down the concrete alternatives to standard autoregressive transformers—linear attention hybrids, Gated DeltaNet, text diffusion, and recursive small models.
articleSebastian Raschka, PhD
Understanding the 4 Main Approaches to LLM Evaluation (From Scratch)
An educational article that explains the four primary methods for evaluating large language models — multiple-choice benchmarks, verifiers, leaderboards, and LLM judges.
Why it mattersIt gives a hands-on, from-scratch breakdown of the four main LLM evaluation approaches — including how MMLU scoring and verifier-based evals actually work.
articleSebastian Raschka, PhD
Understanding and Implementing Qwen3 From Scratch
An in-depth educational article that walks through the Qwen3 LLM architecture and reimplements it from scratch in pure PyTorch.
Why it mattersIf you want to actually understand how modern open-weight LLMs work, this walks you through reimplementing Qwen3's dense and Mixture-of-Experts variants in pure PyTorch.
articleSebastian Raschka, PhD
From GPT-2 to gpt-oss: Analyzing the Architectural Advances
An in-depth technical article breaking down the architectural design choices in OpenAI's gpt-oss-120b and gpt-oss-20b open-weight models, comparing them against GPT-2 and Qwen3.
Why it mattersIf you're evaluating or building on OpenAI's gpt-oss models, this walks through their exact architectural choices (MoE, sliding-window attention, MXFP4 quantization) with clear comparisons to GPT-2 and Qwen3.
articleSebastian Raschka, PhD
The Big LLM Architecture Comparison
An in-depth technical article that compares the architectural designs of modern flagship open-weight LLMs, from DeepSeek-V3 and OLMo 2 to Kimi K2 and Gemma.
Why it mattersIf you're building on or evaluating open-weight LLMs, this gives you side-by-side architectural breakdowns (MLA vs GQA, MoE routing variants, normalization placement) with code.
articleSebastian Raschka, PhD
LLM Research Papers: The 2025 List
A topic-organized collection of 200+ large language model research papers from the first half of 2025, curated by Sebastian Raschka.
Why it mattersA single, thematically organized reference to 200+ of the most important 2025 LLM papers — grouped by reasoning, RL, and multimodal themes.
articleSebastian Raschka, PhD
Understanding and Coding the KV Cache in LLMs from Scratch
A hands-on technical tutorial that explains how key-value (KV) caches speed up LLM inference, walking through a from-scratch, human-readable PyTorch implementation.
Why it mattersIf you want to actually understand how KV caching accelerates LLM inference rather than just calling an API, this walks through a human-readable from-scratch PyTorch implementation and demonstrates a concrete ~5x speedup on a 124M model.
articleSebastian Raschka, PhD
Coding LLMs from the Ground Up
A ~15-hour video course by Sebastian Raschka that teaches you how to build a large language model from scratch in Python.
Why it mattersIf you want to understand how LLMs work under the hood rather than just calling an API, this ~15-hour course walks you through building one in PyTorch end-to-end — tokenization, attention, pretraining, and finetuning.
articleSebastian Raschka, PhD
Common pitfalls when building generative AI applications
The mistakes that recur when teams build with foundation models, starting with the most common one: reaching for generative AI where it was never the right tool.
Why it mattersChip Huyen catalogs the failure modes she's watched teams hit repeatedly—starting with the big one: reaching for generative AI when a simpler solution would do.
articleChip Huyen
Agents
A treatment of agents in the classical sense, rational agents acting in an environment, and what foundation models change about which agentic applications are finally buildable.
Why it mattersA rigorous, literature-grounded framework for defining, building, and evaluating AI agents—covering tools, planning, and the new failure modes agents introduce—rather than a surface-level overview.
articleChip Huyen
Building A Generative AI Platform
The components that recur across generative AI platforms once you look at how companies actually deploy them, built up from the simplest possible architecture rather than presented as a finished diagram.
Why it mattersA clear, incrementally-built reference architecture for production genAI systems — showing when and why to add RAG, guardrails, model gateways, caching, and orchestration.
articleChip Huyen
What I learned from looking at 900 most popular open source AI tools
What a survey of the 900 most popular open-source AI repositories shows about how the ecosystem has changed, four years after the same exercise on the open ML landscape.
Why it mattersA structured, data-driven map of the open source AI tooling landscape — organized into infrastructure, model development, and application layers.
articleChip Huyen
Predictive Human Preference: From Model Ranking to Model Routing
Instead of ranking models and picking one, predict which model a user would prefer for a specific prompt, turning human preference data from a leaderboard into a routing signal.
Why it mattersIf you're deciding which LLM to call for each prompt, this lays out how to predict which model a user will prefer per-query and route accordingly.
articleChip Huyen
An index of the vibe-coding frontier. Corrections welcome.