Following the right people on X and LinkedIn is one of the highest-leverage information habits for anyone working with AI in 2026. The pace of model releases, research breakthroughs, and policy shifts is fast enough that any single news source falls behind quickly. Aggregating signal from a curated list of practitioners, researchers, founders, and policymakers is now standard practice.
This is the AITraining2U Editorial Team's curated list — expanded to 60 people in 2026 to give proper coverage of China's frontier AI labs (DeepSeek, Zhipu, Moonshot, MiniMax, Alibaba/Qwen, ByteDance, Baidu, Baichuan, Tsinghua), which are now reshaping the global model landscape. The selection deliberately spans frontier labs, academic research, applied practice, education, investment, ethics, and the major Chinese AI ecosystem.
Each entry includes a verified photo (Wikipedia where available, initials avatar otherwise), the person's primary affiliation, why they are worth following in 2026, and direct links to X (Twitter), LinkedIn, and personal/company website where each is publicly available.
Frontier lab founders & product leaders
Sam Altman
Co-founder & CEO, OpenAI
The most public face of frontier AI. Posts on AI strategy, GPT roadmap, infrastructure deals (Stargate), and the macro-economics of compute. His threads are the closest thing the industry has to a real-time barometer for OpenAI's direction and the broader frontier race.
Daniela Amodei
Co-founder & President, Anthropic
Anthropic's product, policy, and operations leadership. Less prolific publicly than Dario but increasingly the voice on enterprise deployment, responsible scaling policy, and the practical mechanics of running a frontier lab.
Demis Hassabis
Co-founder & CEO, Google DeepMind; 2024 Nobel Laureate (Chemistry)
Won the 2024 Nobel Prize in Chemistry for AlphaFold. Posts about scientific AI applications, the path to AGI, and DeepMind research milestones. The most credentialed academic voice running a frontier lab in 2026.
Mark Zuckerberg
Founder & CEO, Meta Platforms
Meta's open-weight Llama strategy, AR/VR convergence, and AI-in-WhatsApp/Instagram are reshaping consumer AI. Worth following for the open-weights perspective specifically — Meta has been the largest enterprise sponsor of openly-released frontier models.
AI researchers & scientists
Geoffrey Hinton
Turing Award (2018); 2024 Nobel Laureate (Physics); ex-Google
Foundational researcher in deep learning, now spending public attention on AI safety and existential risk. Won the 2024 Nobel Prize in Physics. Posts and interviews are some of the most weighted voices in the safety conversation.
Yoshua Bengio
Turing Award (2018); Founder, Mila; Professor at Université de Montréal
The third Turing Award godfather. Leads Mila, one of the world's top deep learning research institutes. Increasingly active on AI safety and governance, including the International AI Safety Report.
Andrew Ng
Founder, DeepLearning.AI & Coursera; Adjunct Professor, Stanford
Co-founded Coursera and Google Brain. The world's most influential AI educator — his weekly newsletter The Batch is required reading for practitioners. Posts pragmatic enterprise AI advice that lands without hype.
Fei-Fei Li
Stanford HAI co-director; co-founder, World Labs; ex-Google Cloud Chief Scientist
Created ImageNet, which catalysed the deep learning revolution. Co-founded Stanford's Human-Centered AI Institute and World Labs (spatial intelligence). Posts straddle research, policy, and the human dimension of AI.
François Chollet
Founder, Ndea; creator of Keras; author of "Deep Learning with Python"
Created Keras, one of the most widely used deep-learning frameworks. Designs ARC-AGI, a benchmark for general intelligence. The clearest sceptical-but-rigorous voice on what current LLMs can and cannot do.
Tri Dao
Chief Scientist, Together AI; Assistant Professor, Princeton; creator of FlashAttention
Authored FlashAttention, the algorithm that made long-context LLMs practical. Now leads systems research at Together AI. Posts on inference efficiency and long-context architectures.
Christopher Manning
Director, Stanford AI Lab; Professor of Linguistics & Computer Science, Stanford
NLP pioneer; advisor to many of the most cited researchers in modern language modelling. Long-form writing and lectures are some of the best entry points for understanding why current language models work.
Educators & AI communicators
Practitioners & open-source builders
Investors & ecosystem leaders
Policy, ethics & safety voices
China AI lab leaders
Liang Wenfeng
Founder & CEO, DeepSeek; co-founder, High-Flyer
Personally funded DeepSeek through his quantitative hedge fund High-Flyer Capital Management. Ranked #5 on AI Magazine's Top 100 AI Leaders 2026 after DeepSeek-R1's open-weight release reshaped the cost economics of frontier reasoning. Rarely posts publicly — follow DeepSeek's research releases instead.
Yang Zhilin
Co-founder & CEO, Moonshot AI (Kimi)
Founded Moonshot AI in March 2023 in Beijing. Kimi is among the most-used AI assistants in China and the most prominent challenger to ByteDance's Doubao in consumer chat. Researcher at Tsinghua and CMU before founding Moonshot.
Tang Jie
Co-founder & Chief Scientist, Zhipu AI; Professor, Tsinghua University
Spun out from Tsinghua University in 2019. Chief scientist behind Zhipu's GLM model family. Was a technical lead in BAAI's Wudao project — many of the researchers who went on to found other Chinese AI labs (including Moonshot's Yang Zhilin) trained under him.
Zhang Peng
CEO, Zhipu AI
Public face of Zhipu AI, one of China's "AI Tigers" and the highest-profile Tsinghua spin-out. Frequently the spokesperson for Chinese open-weight model releases. Zhipu went public in 2026.
Yan Junjie
Founder & CEO, MiniMax
Founded MiniMax in 2021. The Shanghai-based lab now valued at $3B+ produces frontier multimodal models including Hailuo (video) and the MiniMax-01 family with 4M-token context. Went public in 2026.
Eddie Wu
CEO, Alibaba Group; oversees Qwen strategy
Took over as Alibaba CEO in late 2023 with an explicit AI-first mandate. The Qwen open-weight model family — the third most-used AI assistant globally behind ChatGPT and Doubao in 2026 — sits under his strategic direction.
Robin Li
Co-founder & CEO, Baidu; pioneer of Chinese internet search
Founded Baidu in 2000. Among the earliest Chinese tech leaders to pivot publicly to AI as core strategy, with the Ernie (Wenxin Yiyan) model family. Frequently sets the public tone of Chinese AI commercialisation debates.
Wang Xiaochuan
Founder & CEO, Baichuan AI; ex-CEO, Sogou
Founded Baichuan AI in 2023 after stepping down from Sogou (acquired by Tencent). Baichuan is one of the "AI Tigers" producing the Baichuan open-weight model series. Strong technical pedigree and a pragmatic enterprise focus.
Lin Junyang
ex-Tech Lead, Alibaba Qwen (departed March 2026)
Was the technical lead and public face of Alibaba's Qwen models for several years. Announced model releases, shared benchmark results, and engaged the global developer community. Departed Alibaba in early 2026 — watch where he lands next.
Liang Rubo
CEO, ByteDance; oversees Doubao
Took over as ByteDance CEO in 2021. ByteDance's Doubao is the most popular AI chatbot in China by usage. Liang shapes ByteDance's broader AI strategy spanning consumer assistants, search (Doubao Search), and enterprise products.
Andrew Yao
Turing Award (2000); Dean, Institute for Interdisciplinary Information Sciences, Tsinghua
The first Chinese-born Turing Award winner. Founded Tsinghua's elite Yao Class which has produced many of China's frontier AI researchers. Frequently regarded as the godfather of theoretical computer science in China and the academic root of much of the country's AI talent pipeline.
Daya Guo
Lead researcher on DeepSeek-R1; DeepSeek-AI
Lead author on the DeepSeek-R1 paper that demonstrated reasoning could emerge from pure reinforcement learning. The arXiv paper became one of the most-cited AI works of 2025 and reshaped the open-weight reasoning landscape.
How to use this list
Follow 10–15 of these to start, picked across categories. Aim for breadth over depth: a US founder, a Chinese lab leader, a researcher, an educator, a practitioner, a policy voice. The goal is to triangulate — when three different people from three different categories reference the same model release, paper, or shift, that is the signal worth paying attention to.
The China-watching layer matters more than ever in 2026. DeepSeek, Qwen, Zhipu, MiniMax, Moonshot and Doubao have collectively reshaped the frontier-model landscape, particularly on the open-weight side. Following at least 3–4 of the Group G figures is the difference between knowing what OpenAI did last week and understanding the global model conversation.
Refresh the list every 6–12 months. The names change. Some figures leave their roles (Mira Murati, Ilya Sutskever both left OpenAI in 2024; Lin Junyang departed Alibaba in March 2026); others rise rapidly into new prominence. The discipline of following an intentionally curated list rather than whatever the algorithm serves you is the durable habit.
If you want to keep up with Malaysian-specific AI insights and our practitioner perspectives, follow AITraining2U on LinkedIn.
