Founder of 海国图智 Intellisia Institute, an independent think tank on international affairs in China. Professor at Jinan University, China.
Guangzhou, China
Joined March 2012
- Tweets8.6K
- Following5.4K
- Followers9.8K
- Likes391
Chen Dingding retweeted
马斯克在最新的访谈里聊到 AI 抢走所有工作时，用了一个极其温柔、但也极其残酷的比喻。
他说：未来人类的工作，大概率会变成园艺。
你根本不需要自己在后院种番茄和土豆，超市里买到的永远比你自己种的更快、更便宜、品相更稳定。
但今天依然有很多人在阳台种花、在院子里拔草，甚至花一整个下午为朋友做一顿自己种出来的晚餐。
这件事在纯粹的工业效率上毫无意义，但它带着一种人类之间独有的温度。
很多人聊到 AI 替代人类，第一反应永远是恐慌、失业、降薪、社会阶层崩塌。
但马斯克推演到了一个完全反直觉的终局：
当 AI 和具身智能把人类在生产力上的最后一滴汗水都彻底碾压之后，劳动的定义本身被击碎了。
过去几千年，人类是通过有用性来确立自我价值的。
你能搬砖、能算账、能写代码、能做报表，你被按月支付薪水，你的尊严建立在社会需要你干活的基础上。
但当 AI 在任何一项具体任务上都比人类做得更好、成本更低时，有用性就不再是稀缺品了。
这时候人类不会灭绝，人类会集体进入一种被迫的退休状态。
就像那些退了休却过得极度充实的老人一样，他们每天早起练字、下棋、养花、钓鱼，没有一件事情是在给社会创造 GDP，但每个人都活得自得其乐。
未来所有的实用性工作，都会被自动化彻底吃掉。
而那些效率最低、最耗时间、最不划算的非标动作，反而会成为人类精神生活里最后的奢侈品。
写一封手写信、亲手烤一块面包、花三个月陪孩子搭一个歪歪扭扭的树屋，这些事情的价值恰恰在于它们是低效的、笨拙的，并且充满了人类自己的真实痕迹。
机器负责解决生存的效率，人类退回生命本身的体验。
所以这里有一个很重要的问题值得我们思考，如果未来你做任何事情都不再能换来超越机器的报酬，你今天真正该想清楚的是：
剥离掉靠打工换来的社会身份之后，你到底靠什么让自己感到快乐？
掌舵红杉资本二十年的 Doug Leone，在访谈里说了句几乎得罪所有同行的话：
风投这行犯过最大的错误，其实不是投错了哪家烂公司，而是太早卖掉了真正的赢家。
他直接摊开了红杉历史上最值钱、也最痛的四笔减持：
红杉曾经持有英伟达 20% 的股份，卖掉了。
他们是苹果的第一位机构投资人，卖掉了。
持有过思科四分之一的股份，卖掉了。
持有过 Google 10% 的股份，也卖掉了。
光是 Google 这一家，如果一股没卖，今天红杉手里的持仓就值 5000 亿美元。
风险投资的传统规则，永远是算账面 IRR、保账面回报率、到了退出期赶紧把现金分给 LP。
只要账面赚了 10 倍甚至 50 倍，全行业都会劝你落袋为安，把战绩锁死在基金年报里。
但 Doug Leone 彻底推翻了这个逻辑：
真正伟大的公司，在 IPO 敲钟之后还会复合增长 20 年以上。
只要你稍微有点商业嗅觉，确认这家公司本质上面对的是一个无限市场，你就绝不应该卖出一股。
以前大家以为赚大钱靠的是不断押中新风口，实际上最恐怖的超级复利全在后面。
把 50 亿美元的收益继续留在场上，它能自动滚成 500 亿美元。
传统风投是用 10 年期的封闭式基金，去生搬硬套一个能跑 30 年的无限滚雪球游戏。
为了锁住眼前的账面收益，亲手把人类历史上最值钱的一批资产拱手送给了二级市场。
这对我们做业务、做技术和做积累也是同一种提醒，
最难的从来不是最初挑对了好方向，而是当好资产开始给你爆发式反馈的时候，克制住自己想要落袋为安的小聪明。
平庸的东西才需要频繁波段，真正能穿越周期的飞轮，拿住就是全部的策略。
当然这个逻辑之所以在红杉身上成立，前提是他们有最顶级的眼光和打破规则的体量，红杉后来甚至特意推翻了成立半个世纪的基金模式做出了永续基金。
对绝大多数人来说，无限市场极其罕见，把平庸资产误当成无限赢家死扛，是更致命的陷阱，
而赚大钱靠的也不是频繁兑现的精明，关键在无限市场里按住自己想套现的手。
Chen Dingding retweeted
“China's power needs continue to grow by roughly one Germany per year.”
Vast majority of this growth still comes from regular industry rather than new drivers like EVs or data centers.
Lots of great data in these slides.
In power generation terms, the sector has almost doubled over the last decade and will crack 11,000 TWh by the end of this year (or 11 Petawatt hours). That's almost roughly 2.5x more than the next-largest country (the USA). China's power needs continue to grow by roughly one Germany per year.
Although thermal generation has continued to rise over the last decade, it's visually clear from this sequence that the lion's share of growth over the last decade, especially in the last few years, was met by non-thermal sources.
Chen Dingding retweeted
🚀 Introducing DeepInfer, our in-house inference framework, now fully open source!
Powering DeepSeek-V4.1-Flash at scale with high throughput and low latency. We’re sharing our production stack to help the community lower compute costs, reduce energy use, and make AI more affordable.
🔗 github.com/deepseek-ai/DeepI…
1/4
Chen Dingding retweeted
Alexandr Wang — Scale AI co-founder, now Meta's Chief AI Officer — made a point most Western AI labs don't want to sit with:
China is ahead on AI data. Not in model architecture. In the infrastructure that feeds models.
The mechanism matters here: a decade of systematic digital infrastructure build-out means China has data throughput and cross-sector integration baked into the economy at a structural level. Wang's framing — data as "the new oil of AI" — isn't about raw volume. It's about collection velocity and scenario diversity. Those two things are much harder to replicate than compute.
The US leads on chips and model research. But if training data becomes the binding constraint for the next generation of capable models, the advantage flips. That's not a dramatic prediction. It's the logical consequence of how these systems scale.
Worth noting who's saying it: Wang runs the company that builds data pipelines for the biggest AI labs on earth. He knows where the bottlenecks are better than anyone.
Chen Dingding retweeted
FirstFT: China’s economy picks up pace as AI boom supercharges exports ft.trib.al/v8EfB5u
Chen Dingding retweeted
The asymmetry is shifting. Washington holds leverage over advanced semiconductors, SME, and EDA and other key software. Beijing holds leverage over rare earths, magnets, materials processing, industrial equipment, and, increasingly, technology and IP that others want.
Chen Dingding retweeted
Beijing’s latest export controls issued via MOFCOM announcements are yet another step in China's 7-year effort to build a full counter-toolkit to Washington’s weaponization of tech supply chains. What began with ZTE and Huawei is now a much broader contest over who controls access to critical technologies. Beijing has learned relatively fast. A short thread.🧵
Chen Dingding retweeted
✨🇨🇳China’s self-driving vending vehicle automatically returns to its charging parking spot to recharge after a day of work.
Chen Dingding retweeted
🚀 DeepSeek-V4.1-Flash-1017 is here!
Stronger benchmark performance, faster task completion, and shorter outputs than the previous version—a Pareto improvement in performance and cost. Now live on our website, app, and API.
🛠️ Optimized for @DeepSeekHarness, Tencent CodeBuddy, WorkBuddy, and more. Upgrade to DeepSeek Harness v0.2.1-rc.1 or later for the best experience with the new model.
1/4
Chen Dingding retweeted
Beijing has issued its first dedicated joint Central Committee/State Council document on “new quality productive forces” (新质生产力): 7 sections, 19 articles. It turns a slogan coined by Xi in 2023 into a governing framework, and several changes from earlier formulations stand out. There is a lot going on here including around AI. A short thread 🧵 news.cn/politics/20261009/37…
Chen Dingding retweeted
Both China and the US are feeding an AI investment bubble – and Shenzhen is the place to talk about it. buff.ly/1mv9dLx
What the US and China Should Discuss at APEC thediplomat.com/2026/10/what…
Next month in Shenzhen, AI should be the focus between Xi and Trump.
Chen Dingding retweeted
Cruising through China at 350 km/h on high-speed rail. You really have to see China to appreciate the incredible development here. "Catching up" belongs to the past...
** Napoleon allegedly (though doubtfully) said, “Let China sleep, for when she wakes, she will shake the world”
Chen Dingding retweeted
In Beijing, meeting with my counterpart, Minister Wang Wentao. Building on three intensive months of engagement, we're working hard to rebalance EU-China trade. It's crucial that the first phase of our talks under the Trade and Investment Consultations delivers tangible outcomes.
Chen Dingding retweeted
Interesting. Beijing seems to want direct discussions between US and Chinese tech firms on AI:
“The sources also said there were initial discussions to bring together executives from top Chinese and American tech firms to discuss artificial intelligence”
Exclusive | Beijing wanted more from Xi’s Washington state visit
scmp.com/news/china/diplomac…
Chen Dingding retweeted
.@SusanShirk1 tells @bcamcrane that the United States should “undertake a more pragmatic debate and reassessment of [its] approach to China.”
The run-up to the 2028 presidential election will provide a critical opportunity to do so.
thewirechina.com/2026/09/13/…
Chen Dingding retweeted
1/ On geopolitics through the end of this year, I sense four major storylines: 1) Iran; 2) Ukraine; 3) impact of US midterms on Trump's foreign policy; and 4) China-Europe trade dynamics. Of these four, the China-Europe trade dynamic has widest range of trajectories. (short 🧵).
Chen Dingding retweeted
China is not just exporting more of everything.
China’s share of global exports has been declining in categories such as finished clothing, shoes, and bags.
