15 AI Accounts You Need to Follow on X (Twitter) in 2026

Most of what you see about AI on LinkedIn today was already discussed on X (Twitter) two months ago. If you're serious about staying ahead in AI — whether you're building, researching, or just trying to keep up — your X feed is your most important tool.

But here's the problem: 90% of AI Twitter is noise. Hype threads, engagement bait, and recycled takes. The signal lives with a small set of people who are actually doing the work — publishing research, building tools, and pushing the field forward.

Here are the 15 accounts that consistently deliver real value. Inspired by @FourVork's viral thread and curated with additional context on why each account matters.

The Pioneers

1. Andrej Karpathy — @karpathy

Former Director of AI at Tesla, founding member at OpenAI, and the man who coined "vibe coding." Karpathy's tweets don't just comment on AI — they create the narratives you'll see on LinkedIn two months later. His Zero to Hero YouTube series remains one of the best resources for understanding LLMs from scratch. When Karpathy tweets, the industry listens.

2. François Chollet — @fchollet

Creator of Keras and the mind behind ARC-AGI, the benchmark designed to measure genuine intelligence rather than pattern matching. Chollet posts thoughtful, often contrarian takes on what intelligence actually means, why current benchmarks are broken, and where AI's real limitations lie. If you want to think deeper about AI beyond the hype cycle, Chollet is essential.

3. Yann LeCun — @ylecun

Turing Award winner, deep learning pioneer, and Meta's Chief AI Scientist. LeCun is known for big-picture research takes, sharp critiques of AGI doomerism, and yes — the occasional drama. Love him or argue with him, his posts force you to think about the fundamental architecture of intelligence.

The Educators

4. Andrew Ng — @AndrewYNg

The godfather of accessible AI education. Co-founder of Coursera, creator of deeplearning.ai, and the person who has probably taught more people machine learning than anyone alive. Ng's feed is a mix of practical ML advice, course announcements, and real-world implementation insights. If you're learning, he's your north star.

5. Sebastian Raschka — @rasbt

The "build from scratch" guy. Raschka publishes incredibly detailed tutorials on implementing ML and LLM components from the ground up. His book on building LLMs from scratch has become a must-read. If you want to understand what's under the hood rather than just calling APIs, follow Raschka.

6. DAIR.AI — @dair_ai

Weekly ML and AI paper threads that make cutting-edge research actually accessible. DAIR.AI curates and explains the papers that matter, saving you hours of sifting through arXiv. High-signal, consistently updated, and one of the best ways to stay current without drowning in academic jargon.

The Deep Divers

7. Lilian Weng — @lilianweng

Ex-OpenAI researcher known for her legendary "Lil'Log" blog posts — the kind of deep, comprehensive write-ups that become reference material for the entire field. Her threads break down complex LLM research into structured, digestible explanations. If a topic is confusing, check if Lilian has written about it first.

8. Jeremy Howard — @jeremyphoward

Co-founder of fast.ai and a tireless advocate for democratizing deep learning. Howard believes anyone can learn ML without a PhD, and he's built the tools and courses to prove it. His takes on AI news are interesting, honest, and occasionally spicy. A builder-educator hybrid at his best.

9. Simon Willison — @simonw

Django co-creator turned LLM tool builder. Simon doesn't just talk about AI — he builds with it, publicly, in real time. His experiments with prompting, his LLM CLI tools, and his engineering breakdowns are gold for anyone who wants to go beyond theory and actually ship things with AI.

The Curators & Thinkers

10. AK — @_akhaliq

The single best account for never missing an important AI release. AK curates the latest arXiv papers, model drops, and open-source AI releases with machine-like consistency. If a new model, paper, or tool drops anywhere in the world, AK has already posted about it.

11. John Carmack — @ID_AA_Carmack

The legendary game developer behind DOOM and Quake, now fully focused on AGI research. Carmack's takes blend low-level systems thinking with high-level AGI philosophy in a way nobody else can. His posts make you reconsider the problem from first principles.

12. Gwern — @gwern

The anonymous polymath behind some of the highest-quality long-form AI research writing on the internet. Gwern's essays and research notes go deeper than almost anyone — on scaling laws, AI capabilities, forecasting, and more. If you want rabbit holes that actually teach you something, start here.

The Visionaries

13. Riley Goodside — @goodside

The prompt engineering pioneer. Goodside was testing LLM capabilities, finding edge cases, and breaking models before most people even knew what prompt engineering was. His evaluations and experiments reveal what AI models can actually do — and where they convincingly fail.

14. Fei-Fei Li — @drfeifei

The computer vision pioneer who created ImageNet — the dataset that arguably kicked off the modern deep learning revolution. Now focused on human-centered AI and spatial intelligence research at Stanford. Li brings a perspective that's rare in the field: technical depth combined with genuine concern for how AI impacts humanity.

15. Demis Hassabis — @demishassabis

CEO of Google DeepMind, Nobel Prize winner, and the mind behind AlphaFold, AlphaGo, and Gemini. Hassabis represents what happens when you combine deep scientific ambition with world-class execution. Following his work gives you a window into the frontier of what AI can achieve when pointed at real scientific problems.

How to Actually Use This List

Don't just follow all 15 and forget about it. Here's what actually works:

Turn on notifications for your top 5. Pick the 5 accounts most relevant to your work and enable notifications. What they tweet today becomes the mainstream conversation in weeks.

Create a private list. Add all 15 to a dedicated X list so you can check it without algorithm interference. Pure signal, zero noise.

Engage, don't just consume. Reply thoughtfully, quote tweet with your own take, build in public alongside these people. The AI community on X rewards contribution over consumption.

Who Did We Miss?

This list is a starting point, not the final word. The AI space moves fast and new voices emerge constantly. If you think someone deserves a spot, drop a comment — the best lists are built by the community.

Originally inspired by this thread from @FourVork, which pulled 256K+ views and counting.