Co-founder @ MatchPoint & AGI Villa · Head of City Ecosystem @datawhale2018 · @n8n_io Ambassador · Host of Re:Organize Exploring AI, people & organizations 🌍
ChiangMai
Joined January 2026
- Tweets45
- Following190
- Followers19
- Likes137
Today an agent spent almost two hours on local screenshots / a walkthrough before opening a PR.
Process isn't a dead line. If a step isn't required for the result, skip it — don't dry-wait.
Worse: agents often treat optional checkpoints as must-haves, then silently wait to finish that node before reporting. If you don't chase them, they can stay stuck forever without telling you where.
Muse Wonder just dropped.
Running ~30 bots day to day, the unlock for me wasn't another model race.
It was: can the bot prove it finished, and did I skill-ify the ops so I don't babysit the same failure twice.
Running a team of agents, memory is the wall. A few of them open a browser and the desktop runs out of RAM, so the task fails. Anyone else hit this, and what actually fixed it?
ask your Grok @Bot to monitor X for user feedback on your products! send feature requests to your issue tracker, bug reports to a cursor cloud agent or project to triage and fix
the loop is complete
The iron-flower show in Situ town holds a Guinness record, and it really is stunning in person. – at People's Republic of China
Spent the National Day holiday in southeast Shanxi (Changzhi + Jincheng), driving the Taihang No. 1 Road along South Taihang.
Longmen Temple alone has buildings from the Five Dynasties, Song, Jin, Yuan, Ming and Qing. In one old temple there's a tree that has lived 1,400+ years. Standing in front of it, you just feel awe.
Spent the National Day holiday in southeast Shanxi (Changzhi + Jincheng), driving the Taihang No. 1 Road along South Taihang.
Longmen Temple alone has buildings from the Five Dynasties, Song, Jin, Yuan, Ming and Qing. In one old temple there's a tree that has lived 1,400+ years. Standing in front of it, you just feel awe. – at People's Republic of China
APSARA Conference · AgentCore forum: we co-released the AI Agent Handbook, and I hosted a roundtable.
Talking with the guests — enterprises aren’t really asking how to *build* Agents anymore. It’s continuity, scale in real ops, and how humans and Agents work together.
As personal Agents show up, personal knowledge bases and your own data sedimentation matter more — and more people will feel both the pain and the charm of working with several Agents at once.
At Ant International today with 40+ women founders from around the world — talking AI.
Lots of interest in getting AI into real work. Most use I heard is still day-to-day; putting it into the actual business is still early exploration for many people here.
When I asked about FDE (Forward Deployed Engineer), most people in the room weren’t sure what it actually does. That gap itself felt worth noticing.
How are you seeing FDE from where you sit?
Glad we got to exchange. Looking forward to more collaboration from here.
Tried Grok Bot yesterday. This “personal agent” wave finally feels real — not another chat box.
You talk to it like a person. Then it takes over and finishes the next steps.
Yesterday afternoon I ask an agent and asked it to order milk tea on Ele.me.
Halfway through: the pay button was blocked by page zoom, and I couldn’t drive the browser. It asked to take over for a few seconds, fix the scale, then hand control back so I could pay. Auth + location permissions ate a chunk of time in the middle.
From zero, all the way into LLM inference and AI Infra. Let's go! 🚀 – at Taiwan
SGLang × Datawhale present zero-to-sglang, the official AI Infra course. Class is now in session! 🎉
SGLang is one of the mainstream open-source high-performance inference engines for large models, serving everything from a single GPU to large clusters with low latency and high throughput.
Founded in 2018, Datawhale is an AI open-source community gathering explorers to share advanced AI knowledge. Core value: grow with learners, for learners.
In this course you will:
Start by reading code, then build a mini-sglang with your own hands
Go deep into the real SGLang source: attention backends, CUDA Graph, prefill-decode disaggregation, and more
Learn how open-source collaboration works and land your first high-quality PR
Swap hands-on experience with the community and round out your AI Infra knowledge
New chapters land continuously. Search 🔍 "zero-to-sglang" to check out the course, and tell us in the comments what you most want to learn.
From zero, all the way into LLM inference and AI Infra. Let's go! 🚀
#SGLang #Datawhale #AIInfra #LLMInference #LLMServing #OpenSource #OpenCourse #OpenSourceCommunity #AILearning
Darren Su retweeted
📘 Introducing the Zhihu AI Cookbook (CN)
Saved dozens of AI articles but still not sure where to begin? We’ve organized some of Zhihu’s best technical discussions into clear, structured learning paths.
Over the past few years, Zhihu has become home to a wealth of in-depth AI writing from researchers, engineers and practitioners. The Zhihu AI Cookbook brings together some of the most useful and enduring pieces in one place.
The first edition brings together 80+ selected posts from 68 contributors across three tracks:
🖥️ AI Infra
🧠 Reinforcement Learning for LLMs
🤖 Embodied AI
It’s not another textbook. Think of it as a map connecting explainers, source-code walkthroughs, project retrospectives, interview insights and open-source tutorials.
🧭 How to use it:
1️⃣ Start with the introductory guide to understand the key concepts and prerequisites.
2️⃣ Choose the chapter that matches what you need right now.
3️⃣ Read each selection with a guiding question, then use the self-check prompts to test your understanding.
4️⃣ Follow the links to the original Zhihu posts for full derivations, implementation details and community discussions.
5️⃣ Put the ideas into practice with hands-on projects and open-source tutorials.
Whether you’re exploring a new research direction, working through a project or preparing for technical interviews, the Zhihu AI Cookbook can help you find the right place to start and build a more complete understanding.
🤝 With open-source tutorials contributed by our content partner, @datawhale2018.
🌏 This edition is currently available in Chinese. An English version is coming soon.
🚀This is only the first edition. More topics and learning paths are on the way…
🔗 Explore the Zhihu AI Cookbook (CN):
my.feishu.cn/wiki/space/7685…
💻 View it on GitHub:
github.com/Zhihufrontier/zhi…
If you find it useful, give the repo a star and let us know what topic we should cover next. 🌟
#AI #LLM #EmbodiedAI #AIInfrastructure #ReinforcementLearning #MachineLearning
Great conversations at Google I/O Connect {China} with @Paul_Kinlan , @rickyrobinett , and @dsmcl .
Chatted with Paul about Chrome + WebMCP, and asked Ricky & David what matters most in developer relations. – at People's Republic of China
Darren Su retweeted
🤖 When everyone has an AI assistant, what’s next for developers?
AI Agents are evolving beyond simple conversations — moving toward task execution, workflow collaboration, and real-world automation.
In this TinTinAI roundtable, we’re bringing together experts from 0G, AGI Villa, TeamWiseflow, and leading AI Agent developers to explore:
✅ What makes AI Agents different from traditional AI tools?
✅ Which real-world use cases are emerging first?
✅ How can developers seize new opportunities in the Agent era?
✅ Could Web3 become a key infrastructure layer for the Agent Economy?
📅 July 8, 2026 | 20:00 (UTC+8)
📍 Watch Live: TinTinAI WeChat Video Channel
📌 Registration: Scan the QR code on the poster to reserve your spot
Speakers:
🎙 @jtsong2｜Head of APAC, @0G_labs
🎙 Jolene｜Software Development Engineer
🎙 Darren @zenshipai｜Co-founder, AGI Villa
🎙 Zeming Zhao @Shanghaibi31603｜Creator of TeamWiseflow, Zhihu Contributor
#TinTinAI #Web3
Darren Su retweeted
SuperAI is heading to Shanghai.
An invite-only builders mixer with @01AI_Yi ,@Google, @StepFun_ai, @WAVE599395 @SGglobalnetwork, AGI Villa and XNODE.
Bringing together founders, operators, and investors from China's AI ecosystem and beyond.
