AI & Machine Learning blogs

A community blog on rationality, AI risk, and self-improvement, where members post long essays and discussion.

Top post: Generalizing from one example

Google's official corporate blog, carrying product announcements, AI news, and policy statements from the company.

Top post: Google: G Suite and Gmail to more closely align

Gwern Branwen writes long, heavily researched essays on AI, statistics, nootropics, technology economics, and other subjects.

Top post: Laws of Tech Economics: Commoditize Your Complement

Simon Willison, a developer and co-creator of Django, writes about LLMs, prompt injection, and AI-assisted programming.

Top post: AI agent lethal trifecta: private data, untrusted content, exfiltration vectors

Meta's corporate newsroom publishes announcements on products, layoffs, AI models, research, and policy and content moderation changes.

Top post: A new optional way to share and discover music tv and movies

DeepMind's official blog, announcing research results such as WaveNet, AlphaFold, AlphaCode, and MuZero.

Top post: WaveNet: A Generative Model for Raw Audio

Microsoft's corporate blog, with company announcements on AI, OpenAI, carbon goals, products, and policy commitments.

Top post: Microsoft will be carbon negative by 2030

Daniel Stenberg, lead developer of curl, writes about maintaining the project, security reports, and internet protocols.

Top post: Death by a Thousand Slops

Stephen Wolfram, founder of Wolfram Research, writes long essays on computation, physics, machine learning, ChatGPT, and his personal productivity setup.

Top post: What Is ChatGPT Doing and Why Does It Work? (2023)

Google's research team posts on machine learning, speech, vision, and language models, plus occasional general engineering pieces.

Top post: Nearly all binary searches and mergesorts are broken (2006)

Google Research's blog and publications site, covering ML, systems, and applied research such as camera technology and tabular models.

Top post: Zanzibar: Google’s Consistent, Global Authorization System (2019)

Sam Altman, OpenAI's CEO, writes essays on startups, productivity, and the future of AI and superintelligence.

Top post: Employee Equity

NVIDIA's corporate blog covers GPUs, AI research, gaming hardware, autonomous driving, and data center products.

Top post: GauGAN Turns Doodles into Stunning, Photorealistic Landscapes

Armin Ronacher, creator of Flask, writes about Python, packaging, Unicode, async, and more recently AI agents and tooling.

Top post: Unicode in Python 3

SemiAnalysis is a semiconductor and AI infrastructure research firm covering chips, GPU benchmarks, training costs, and datacenter hardware.

Top post: Google "We Have No Moat, and Neither Does OpenAI" (2023)

Google's official developer blog, announcing AI models like Gemini and Gemma, agent protocols, and changes to developer products.

Top post: Interacting with Gemini through multimodal prompting

Ed Zitron writes critical analysis of the AI industry's finances, focusing on OpenAI, Anthropic, and Big Tech spending and revenue.

Top post: The man who killed Google Search?

Marshall Brain writes about automation, robots, and the economy, and is the author of the science fiction story Manna.

Top post: Manna – AI Utopia or Dystopia? By Marshall Brain

Jeremy Howard and Rachel Thomas run fast.ai, publishing practical deep learning courses and essays on AI, ethics and data science.

Top post: Embracing Swift for Deep Learning

Meta's AI research blog, announcing models such as Llama, V-JEPA and Large Concept Models, plus its MTIA AI accelerator chips.

Top post: Meta Releases Llama 3

Meta's AI research blog announces models, papers and systems such as OPT-175B, LLaMA, ImageBind and Instagram's recommender.

Top post: Meta AI releases large-scale language models with OPT-175B

Google's research blog reports on work in machine learning, language, vision and engineering, including the well-known binary search bug post.

Top post: Nearly all binary search and merge sort implementations are broken (2006)

Databricks' engineering blog covers Apache Spark, data lakehouses, LLM training and inference, and the company's tooling and benchmarks.

Top post: Free Dolly: Introducing the First Truly Open Instruction-Tuned LLM

Andrej Karpathy writes tutorials and essays on neural networks, deep learning, and training models, including a minimal GPT implementation.

Top post: The Unreasonable Effectiveness of Recurrent Neural Networks

Andrej Karpathy's Stanford pages host his deep learning projects, image captioning work and PhD thesis, plus other Stanford CS material.

Top post: ConvNetJS: Deep Learning in your browser

Cursor's blog covers its AI code editor, coding agents, models and benchmarks, written by the company's team.

Top post: Cursor CLI

Google's research division posts on machine learning, translation, quantum chemistry, and algorithm engineering, written by its own researchers and engineers.

Top post: Nearly All Binary Searches and Mergesorts Are Broken (2006)

Distill was an online journal publishing interactive, visual explanations of machine learning research, written by researchers.

Top post: Growing Neural Cellular Automata (2020)

Every is a publication and newsletter network covering AI, business, productivity, and tech, with writers like Dan Shipper and Gareth Edwards.

Top post: Top AI Models Compete in a Game of Diplomacy

Apple's machine learning team publishes research papers and technical write-ups on on-device models, foundation models, speech, and reasoning.

Top post: Deploying Transformers on the Apple Neural Engine

The New York Times engineering team writes about its CMS, HTTPS migration, recipe data extraction, and open source web tools.

Top post: Extracting Structured Data from Recipes Using Conditional Random Fields (2015)

Daniel Miessler, a security professional, writes about infosec, Unix tools, tutorials, and more recently AI.

Top post: Building a Successful Information Security Career (2008)

Nick Bostrom, philosopher at Oxford, publishes papers and essays on existential risk, superintelligence, and the future of humanity.

Top post: The Fable of the Dragon-Tyrant (2005)

Andy Baio, a writer and technologist, covers internet culture, copyright, and AI training data, and created Kind of Bloop.

Top post: Kind of Screwed

The blog for Aider, an open-source AI pair-programming tool for the terminal, covering LLM coding benchmarks and repository mapping.

Top post: GPT-4o takes #1 and #2 on the Aider LLM leaderboards

Gary Marcus, a cognitive scientist and AI critic, writes about the limits of large language models and AI hype.

Top post: The whole thing was a scam

The Machine Intelligence Research Institute publishes research and essays on AI alignment and the risks of advanced artificial intelligence.

Top post: There's No Fire Alarm for Artificial General Intelligence

Sean Goedecke, a software engineer at a large tech company, writes frequently on AI, LLMs, and working in software.

Top post: Seeing like a software company

NVIDIA's developer blog covers GPU architectures, CUDA programming, deep learning and self-driving research, and data analytics.

Top post: Volta: Advanced Data Center GPU

OpenAI's early blog posted research updates on language models, reinforcement learning, Dota 2 bots, and adversarial examples.

Top post: Better Language Models and Their Implications

Mitchell Hashimoto, HashiCorp co-founder and creator of the Ghostty terminal emulator, writes on programming, terminals, Nix, and AI use.

Top post: Ghostty

Company blog of Modular, Chris Lattner's startup behind the Mojo language, covering AI compute, GPU programming, and compilers.

Top post: The Claude C Compiler: What It Reveals About the Future of Software

George Hotz, founder of comma.ai and tiny corp, writes on AI chips, AMD drivers, tinygrad, and technology.

Top post: The tiny corp raised 5M

Terence Tao, a mathematician, posts short notes on Mastodon about mathematics, AI tools, and automated proofs.

Top post: Homepage

Facebook Research's site posts papers, open-source releases, and blog posts on machine learning and other research.

Top post: Prophet: forecasting at scale

Anthropic's interpretability team publishes research papers on reverse engineering how transformer language models work internally.

Top post: The Biology of a Large Language Model

METR is a nonprofit that evaluates AI models' capabilities and risks, including studies of AI's effect on developer productivity.

Top post: Measuring the impact of AI on experienced open-source developer productivity

Edward Z. Yang, a PyTorch developer, writes about Haskell, compilers, and machine learning framework internals.

Top post: How to read Haskell like Python

Chris Stucchio, a data scientist and former high frequency trader, writes about statistics, algorithmic fairness, A/B testing, and trading.

Top post: A High Frequency Trader's Apologia, Part 1

Company blog of Sourcegraph, a code search and intelligence vendor, with product news, Go engineering posts, and AI coding assistant Cody updates.

Top post: Cody answers questions and writes code for you by reading the code graph

The Gradient is an online magazine where AI researchers write about machine learning, NLP, interpretability, and the state of the field.

Top post: Interpretability in Machine Learning: An Overview

Erik Bernhardsson, creator of Annoy and former Spotify engineer, writes on software estimation, hiring, data teams and machine learning.

Top post: Why software projects take longer than you think – a statistical model

AI researcher and co-developer of LSTM networks, who writes on the history of deep learning, creativity theory, and credit for ideas.

Top post: Annotated History of Modern AI and Deep Learning (2022)

Researcher Chris Olah writes illustrated explanations of neural networks, LSTMs, backpropagation, information theory, and representation learning.

Top post: Neural Networks, Manifolds, and Topology

Robin Sloan, a novelist and media inventor, writes about software, AI, writing, and the web.

Top post: An app can be a home-cooked meal (2020)

Cosma Shalizi, a statistics professor at Carnegie Mellon, keeps a long-running personal site of notebooks, reviews and essays on statistics, science and society.

Top post: Review of A New Kind of Science by Stephen Wolfram (2002)

Epoch AI is a research group that publishes data and analysis on AI trends, including compute, costs, energy use and benchmarks.

Top post: How much energy does ChatGPT use?

Timothy Gowers, a Fields medalist mathematician, writes about mathematics, academic publishing, journal reform and AI's role in math research.

Top post: Discrete Analysis – A new type of math journal

Maggie Appleton, a designer and researcher, writes about digital gardens, generative AI, and the web, with finely made visual essays.

Top post: A Brief History and Ethos of the Digital Garden

Matt Mahoney, a data compression researcher, hosts the Large Text Compression Benchmark, compression programs and a book, plus ultramarathon writing.

Top post: Large Text Compression Benchmark