People to Follow (X/Twitter)

X/Twitter is where much of AI moves in real time. Papers get discussed hours before the press notices, and model launches often break here first. Grouped below by role. Handles verified July 2026.

Lab leaders & executives

• Sam Altman ( @sama): OpenAI CEO. Posts product hints and release teasers, plus the occasional cryptic one-liner. 🔄
• Dario Amodei ( @DarioAmodei): Anthropic CEO and co-founder. Long-form essays on scaling, safety, and where capability is headed ("Machines of Loving Grace"); posts sparingly but consequentially.
• Demis Hassabis ( @demishassabis): Google DeepMind CEO. Frontier research and science applications, from AlphaFold successors to Gemini.
• Greg Brockman ( @gdb): OpenAI president and co-founder; infrastructure scale, demos, and engineering culture.
• Clément Delangue ( @ClementDelangue): Hugging Face CEO. Posts on open-source models and the 🤗 ecosystem.
• Yann LeCun ( @ylecun): Turing laureate. Skeptical of LLM hype, focused on world models and self-supervised learning.
• Aravind Srinivas ( @AravSrinivas): Perplexity co-founder and CEO; AI answer-engines and pointed takes on where product value accrues in the stack. 🔄
• Arthur Mensch ( @arthurmensch): Mistral co-founder and CEO. The open-weight, European "sovereign AI" viewpoint.
• Ilya Sutskever ( @ilyasut): co-founder of Safe Superintelligence (SSI), previously OpenAI chief scientist. Rarely posts, but a defining researcher of the era.
• Mira Murati ( @miramurati): co-founder and CEO of Thinking Machines Lab; former OpenAI CTO. One of the most-watched new-lab founders. 🔄
• Mustafa Suleyman ( @mustafasuleyman): Microsoft AI CEO; co-founder of DeepMind and Inflection. Author of The Coming Wave.

Researchers & scientists

• Andrej Karpathy ( @karpathy): ex-OpenAI and Tesla, now at Eureka Labs. Writes explainer threads and "nanochat"/zero-to-hero material on building models from scratch.
• Jim Fan ( @DrJimFan): NVIDIA; embodied agents, robotics, and sim-to-real.
• Sebastian Raschka ( @rasbt): author of Build a Large Language Model (From Scratch). Code-first explainers of LLM internals and training. 🔄
• Nathan Lambert ( @natolambert): Ai2/Interconnects; public analysis of RLHF, post-training, and open-model policy.
• Noam Brown ( @polynoamial): OpenAI. Reasoning, self-play, and test-time compute, the ideas behind thinking models.
• Jason Wei ( @_jasonwei): co-author on chain-of-thought and emergent abilities; short notes on what makes models reason.
• François Chollet ( @fchollet): creator of Keras and the ARC-AGI benchmark; co-founder of Ndea and ARC Prize. A leading skeptic-realist on what counts as reasoning.
• Chris Olah ( @ch402): Anthropic; a pioneer of mechanistic interpretability. The clearest public voice on what's actually happening inside models.

Builders & practitioners (agentic coding is the theme of 2026)

• Peter Steinberger ( @steipete): creator of OpenClaw (the open-source agent that "broke GitHub"), now at OpenAI. Posts on agentic-coding loops and harness design. 🔄
• Boris Cherny ( @bcherny): creator of Claude Code at Anthropic. Shows how the people who build the tools use them, including his "vanilla setup" threads.
• Tibo (Thibault Sottiaux) ( @thsottiaux): OpenAI, ChatGPT/Codex engineering lead. Product updates, rate-limit resets, and roadmap notes. 🔄
• Mario Zechner ( @badlogicgames): creator of pi ("the shitty coding agent"), a minimal, steerable harness. Argues against over-building agents.
• Armin Ronacher ( @mitsuhiko): creator of Flask. Writes long-form on agentic coding (see lucumr.pocoo.org). Also on Mastodon.
• Addy Osmani ( @addyosmani): Google Chrome engineering lead; LLM-coding workflows, "loop engineering," and web performance.
• Theo Browne ( @theo): t3.gg / T3 Chat (YC W22). Takes on AI tooling, TypeScript, and the developer stack (channel in Podcasts & YouTube).
• Simon Willison ( @simonw): Django co-creator, builder of the llm CLI and Datasette. Documents what LLMs can actually do (blog in Blogs & Publications).
• Jeremy Howard ( @jeremyphoward): fast.ai co-founder. Works on accessible deep learning, small-model efficiency, and open science.
• Guillermo Rauch ( @rauchg): Vercel CEO; v0, the AI SDK, and the AI-native web stack that many developers ship on.
• Thorsten Ball ( @thorstenball): Sourcegraph / Amp. Clear writing on how coding agents actually work under the hood.
• Hamel Husain ( @HamelHusain): independent consultant. Evals and LLM-as-judge, the "measure before you ship" discipline that's easy to skip.
• Georgi Gerganov ( @ggerganov): creator of llama.cpp and ggml, the projects that put local LLM inference on everyday hardware. The center of gravity for running models yourself. 🔄

Sharp commentators & analysts

• swyx (Shawn Wang) ( @swyx): coined "AI Engineer," runs Latent Space and the AI Engineer conference. Maps the AI-engineering discipline.
• Andrew Ng ( @AndrewYNg): DeepLearning.AI; education-first commentary; author of The Batch ( Newsletters).
• Pedro Domingos ( @pmddomingos): author of The Master Algorithm. Academic voice skeptical of overclaims.
• Ethan Mollick ( @emollick): Wharton professor; the most-read practical voice on using AI for real work (newsletter in Newsletters, blog in Blogs & Publications).
• Eliezer Yudkowsky ( @ESYudkowsky): the best-known AI-safety and existential-risk voice. Polarizing, but the reference point that debate on the risk side organizes around.

Tip: Build a private X List rather than following people directly. It keeps your main feed manageable. A good starter list: a few builders from the group above, plus Karpathy, Simon Willison, swyx, and one lab leader you follow. If you'd rather not live on X, a developer feed like daily.dev surfaces much of what these accounts share without the doomscroll.