Photo by the editor
# Entry
Artificial intelligence is developing so rapidly that time-honored news outlets and even academic journals often struggle to keep up. More specifically, LLMs see breakthroughs in reasoning, performance, and agentic abilities so often that social media is constantly flooded with them. X (formerly Twitter) continues to be the central hub of the artificial intelligence research community where developers, engineers and researchers can share and exchange ideas in real time.
However, finding high-quality information in the era of algorithmic feeds can be a challenge. To truly benefit from the platform, you need to filter through the noise and find authors who offer deep technical knowledge and actionable insights with the greatest impact. There are some massive, obvious names that everyone probably already follows, so I won’t repeat them here. Instead, this article focuses on accounts that consistently share useful LLM updates, documents, tools, or thoughtful comments. If you want the signal to be better than the noise, these are solid solutions.
# Top 10 X (Twitter) Accounts for LLM Updates
// 1. DAIR.AI (@dair_ai)
// 2. Andrej Karpatia (@karpatia)
Andrzej Karpaty is still one of the best ways to think clearly about deep learning and LLM. When he posts, it’s usually worth reading. He shares intuition, learning tips, and perspectives on where the field is heading. If you care about the basics, this is a must-read.
// 3. Sebastian Raschka (@rasbt)
Sebastian Raschka focuses on implementation and learning by doing. You’ll see tutorials, architecture descriptions, and practical insights into machine learning and LLM. If you actually build models (or want to), his posts are consistently useful.
// 4. alfaXiv (@askalphaxiv)
alphaXiv is powered by arXiv article discovery and discussion, with a community layer for research purposes. It allows you to browse, discuss, and see what other people are using in the latest articles, so you’ll quickly see what’s practical or impactful. Personally, I switched to this over the last month to keep up with trends.
// 5. Artificial Intelligence (@TheRundownAI)
Artificial intelligence destroyed is a large-volume AI news stream that is best used like an ICT service: scan the headlines, click only what is significant, and ignore the rest. Their own position is “the largest AI newsletter”, which fits with how it runs on X – i.e. it’s speedy, broad, and constantly updated. If you want to stay up to date with product launches, financing news and model launches, it does the job.
// 6. AK (@_akhaliq)
AND is one of the most frequently cited accounts for up-to-date arXiv articles, publications of open source models and tools. If something up-to-date comes along, it often appears here quickly. There may be viral content in your feed from time to time, but once discovered, it’s demanding to ignore.
// 7. Ahmad Osman (@TheAhmadOsman)
Ahmad Osman focuses on AI systems, infrastructure and hardware, especially running LLM locally, rather than relying solely on application programming interfaces (APIs). Shares practical insights on graphics processing units (GPUs), inference performance, and self-hosting setups. Honestly, his posts almost convince you to buy a GPU and build your own local LLM setup.
// 8. Matt Wolfe (@mreflow)
Matt Wolfe provides daily AI updates and tool summaries. Very builder affable. If you want to know what up-to-date AI products have launched this week (without checking them out yourself), this account will keep you updated.
// 9. Simon Willison (@simonw)
Simon Willison is excellent for practical operate of LLM. He shares experiments, real tips, tool failures, and forthright reflections on what works and what doesn’t. If you care about actually building with LLMs and not just reading about them, this is one of the best reads below.
// 10. Ethan Mollick (@emollick)
Ethan Mollik talks about the LLM in the context of work, education and real-world impact. Less about the internals of the model, more about “what difference does it make?” If you want thoughtful and original commentary on the impact of artificial intelligence on workplaces and organizations, he has a mighty voice.
# Application
You don’t have to follow hundreds of AI accounts to stay up to date. A petite, well-researched list is usually better. If you care about:
- Tests: DAIR.AI, alfaXiv.
- Deep intuition: Andrzej Karpaty.
- Practical building: Sebastian Raschka, Simon Willison.
- News and tools: Artificial Intelligence Destroyed, Matt Wolfe.
- Systems and infrastructure: Ahmad Osman.
- Work and influence: Ethan Mollik.
Choose based on what you actually want to learn. This alone will cut out most of the noise.
Kanwal Mehreen is a machine learning engineer and technical writer with a deep passion for data science and the intersection of artificial intelligence and medicine. She is co-author of the e-book “Maximizing Productivity with ChatGPT”. As a 2022 Google Generation Scholar for APAC, she promotes diversity and academic excellence. She is also recognized as a Teradata Diversity in Tech Scholar, a Mitacs Globalink Research Scholar, and a Harvard WeCode Scholar. Kanwal is a staunch advocate for change and founded FEMCodes to empower women in STEM fields.