Jielin Dong 董洁林

Scholar on History and Policies of Science and Technology. Former Columnist of Wall Street Journal (Chinese edition). 科技史、科技政策学者，曾任华尔街日报（中文版）专栏作家。

[Jielin Dong 董洁林 @D931931]

AI companies should provide deep educational discounts.

[Fei-Fei Li @drfeifei]

Scientific research is fundamental to advancing civilization and helping people globally to solve the most critical problems, from medicine to materials, from brain science to physics, and much beyond. This is only possible when scientists have access to the best tools of the time to conduct scientific research, including having access to AI-based tools.

[Jielin Dong 董洁林 @D931931]

Agreed. Unfortunately, rules are made for the honest, not the lawbreakers.

[Dario Amodei @DarioAmodei]

Today I'm publishing a new essay, Policy on the AI Exponential. AI is progressing extremely fast—much faster than the policy process was built to handle. The essay lays out where I think the technology is now, and the action needed to close the gap: https://t.co/Lh6PWae178

[Jielin Dong 董洁林 @D931931]

Will Microsoft and Amazon run away from Seattle？

[Zak Kukoff @zck]

Who has the definitive history of how Seattle became crazy far-left despite having Microsoft, Amazon, etc in town?

[Jielin Dong 董洁林 @D931931]

Make sense

[Thierry from arvy 🇨🇭 @ThierryBorgeat]

Citadel Securities just put institutional weight behind what the AI bulls won't say out loud.

In a new macro note titled "Tokenomics," Citadel makes the argument plainly: even the most powerful technology on earth still has to pass through the boring discipline of cost curves, capacity limits, and marginal returns.

The evidence is piling up:
– Amazon removed its token usage leaderboard
– Microsoft cancelled Claude Code subscriptions
– Multiple companies reporting unexpectedly massive token bills

Their conclusion is the part that matters.

Adoption is no longer about what AI can do in principle. It's becoming about the price and scarcity of the inputs needed to run it at scale. Compute. Power. Cooling. Memory bandwidth. Inference budgets. All real, all binding constraints.

And here's the kicker from the chart.

The Silicon Data LLM Token Expenditure Index, a benchmark for how much the market is actually spending on AI tokens, has started rolling over. Citadel reads it as a shift toward cheaper models. Companies substituting away from expensive frontier AI toward "good enough" alternatives.

That's economics 101 doing what it always does. When the price of something rises, people use less of it, or find a cheaper version.

Citadel sees a bifurcation forming. Frontier AI concentrated among a few firms with the balance sheets to absorb the cost. Everyone else quietly downgrading to simpler, cheaper models.

This is the part of every technology revolution the early narrative ignores.

The technology being real was never the question.

The question was always whether the economics could carry the valuations.

When one of the most sophisticated trading firms on earth starts writing about AI in the language of cost curves and rationing instead of limitless demand, the conversation has quietly changed.

The hype was about what AI could do.

The reckoning is about what it costs.

[Jielin Dong 董洁林 @D931931]

Agreed

[Robin Brooks @robin_j_brooks]

There's really no sign - whatsoever - that US inflation is overheating. That has implications for the Dollar. Just imagine what'll happen to inflation when the war ends and oil prices fall. I think markets are wrong to price hikes for the Fed. Very wrong.
https://t.co/jGE8S66IPN

[Jielin Dong 董洁林 @D931931]

一诺的博导居然是CUSPEA，代沟来了

[徒步的骑手 @CaminoTexas]

一诺对谈徒步的骑手：中国・美国，回看25年（下） https://t.co/JF74eQXHaj via @YouTube

[Jielin Dong 董洁林 @D931931]

AI-native!

[Phoenix Yin @Phoenixyin13]

“我在田里雇了一名工程师，它叫 Codex。”
这是AI时代最浪漫且硬核的田野叙事！

最近，这位北海道农民富安弘的故事在全网刷屏。一个毫无背景的前公务员，独自跑到北海道，靠着 ChatGPT 和 Codex，实现了一人管理100公顷农场的硬核神话。

大公司在搞价值千万的智慧农业，而富安弘毅在用几美金的 API 搞平替版赛博农场。来看看他的一键DIY清单：

温室自动通风系统。用 Codex 写代码连接传感器，太热就自动开窗卷帘，甚至能通过 LINE 软件远程发指令。
卫星天眼监控。把卫星数据叠加到农田地图，坐在家里吹着空调，就能一目了然哪块地的西兰花长得不好。
物理外挂。拍一张复杂的配电盘照片直接丢给 AI。

很多人焦虑 AI 会抢走工作，是因为他们把自己定位成了执行者。
但富安弘毅把 AI 当成了24小时不拿薪水的全栈工程师。他没有被技术取代，反而用技术给自己的体力赋能，拿回了对这 1500 亩土地的绝对掌控权。

在这个人人都在卷大模型参数、卷算力、卷 PPT 的时代，这位北海道大叔坐在拖拉机上，用实际行动告诉我们，技术的终点，是脚下踏实的土地。

[Jielin Dong 董洁林 @D931931]

I bought a few tickets to enjoy the game

[Daniel Lurie 丹尼爾·羅偉 @DanielLurie]

San Francisco is ready for the FIFA World Cup! As the Bay Area hosts six FIFA World Cup matches, our city is being activated with watch parties, fan marches, and celebrations in every neighborhood.

Throughout the entire tournament, more than 90 restaurants and bars across the city are hosting World Cup festivities, and venues—including Thrive City, China Basin Park, and Pride House SF—will offer free watch parties for fans.

We’re building on the momentum of Super Bowl 60, which brought $425 million in economic impact to San Francisco, and continuing our focus on public safety and supporting small businesses. Whether you’re visiting from across the globe or celebrating in your favorite neighborhood soccer bar, we’re thrilled to show off the very best of San Francisco.

Fans can go to https://t.co/9D61pnYFKW for everything you need to know about the tournament, including transportation, events, and a new WhatsApp alert service. You’ll find big watch parties and local bars throughout our neighborhoods on the Shop Dine SF World Cup guide at https://t.co/uTeyZBkqNE. 

Let’s go!

[Jielin Dong 董洁林 @D931931]

这里说的是平均每个员工每月花两千美元买AI token.

[Jielin Dong 董洁林 @D931931]

赵长鹏的公司约5千员工，平均每个员工每月两千美元。这比很多旧金山的公司高很多啊。

[GeLun Ding @gelunding]

AI 的增长空间真的有那么大吗？赵长鹏给了我们一些线索。他表示，Binance 每个月在 AI Token 上花费约 1000 万美元，也就是每年 1.2 亿美元。而 Binance 每年的收入约为 175 亿美元。

高盛号称 SpaceX 的 AI 业务收入将在 2030 年达到 3200 亿美元。假设它完全没有竞争对手，仅靠企业客户贡献收入，那么它仍然需要 2667 家像 Binance 一样规模的客户，并且每家都要像 Binance 一样每年花 1.2 亿美元购买 AI Token。

即使 Fortune 500 企业全部按照 Binance 的投入强度使用 AI，一年也只能贡献约 600 亿美元收入（而且其中有一半公司的营收还低于 Binance）。

这意味着，SpaceX 仍然需要额外 2.6 万家年营收 10 亿美元的公司，每年在其 AI Token 上花费 1000 万美元。这简直荒谬。

[Jielin Dong 董洁林 @D931931]

@zhuxia24146714 不是月薪，是每个月每位员工花两千美元买AI token

[Jielin Dong 董洁林 @D931931]

I know that some San Francisco companies stopped hiring interns because of AI and remote work.

[a16z @a16z]

Remote work, not AI, may be the real culprit for recent grad unemployment

Charts of the Week: https://t.co/O5surLTGMJ

[Jielin Dong 董洁林 @D931931]

The AI Paradox: Everyone Wants It, Few Can Afford It

The process of AI-driven enterprise transformation is proving to be far more complex than many expected. For most incumbent companies, there is still little visible evidence that AI is significantly expanding revenue or accelerating sales growth. The costs of adopting AI, however, are immediate and tangible. This raises a fundamental question: Where will the budget for AI investment come from?

First, companies can fund AI initiatives out of their existing profits. But this approach has an immediate impact on financial performance, potentially disappointing shareholders and increasing pressure on management.

Second, if AI substantially improves employee productivity, companies could eventually reduce headcount and redirect the resulting savings toward AI investments. In practice, however, this transition is proving much slower than many anticipated. Most organizations remain in an exploratory phase. Hiring freezes are becoming more common, but the widely predicted wave of large-scale AI-driven layoffs has yet to materialize.

Third, companies can reallocate spending from traditional SaaS software budgets to AI. This shift is already underway, but it is also a gradual process. It is unlikely to free up large amounts of capital quickly, and the total savings available from SaaS consolidation may ultimately be far smaller than the level of investment required for AI adoption.

Against this backdrop, the emergence of AI-native companies deserves close attention. These firms can build their products, services, technology stacks, and organizational structures around AI from the ground up. The critical question is whether they will primarily capture market share from incumbent companies or create entirely new sources of demand and economic value.

Consequently, foundation model companies and AI agent providers need to focus on understanding and improving AI ROI across the entire value chain, optimizing service costs so that the entire ecosystem benefits, rather than blindly pouring capital into ever-larger infrastructure buildouts.

[Jielin Dong 董洁林 @D931931]

AI造成的企业转型过程比想象的要复杂很多。现有企业目前还看不见AI提升其营销规模的迹象，但投入AI的成本是立即的，那么这个预算来自哪里？
1，用公司利润投入AI，这样财务数据立刻受影响，股东不高兴，管理层有压力；
2，如果AI提升了员工效率，可以减少员工，从而腾出预算投入AI，但这个过程比想象的要漫长，各公司基本上处于调研阶段，现在只是停止招新人，而传说中的由于AI而大幅解雇人的过程还没有开始；
3，压缩传统SaaS软件预算、腾出来投入AI，这个过程正在进行中，但这也是个渐进过程，不可能很快释放大笔资金，况且这笔资金的天花板也许比AI所需要小很多。

[Jielin Dong 董洁林 @D931931]

这些企业不是在愚蠢“炫耀”、而是在撒谎。现在几乎没有什么企业靠AI裁掉了2/3员工。

[墓碑科技 @mubeitech]

“千万别到处炫耀，AI 帮你裁掉了三分之二的员工。”
这是 Palantir CEO Alex Karp 刚给硅谷敲的警钟。

现在的科技精英，正在玩一场极其危险的游戏。
为了卷死竞争对手，为了让财报好看。
很多企业疯狂用 AI 替代人工，然后大肆炫耀自己裁掉了大半的员工。

Karp 说，这简直是在自找麻烦。
你这么干，等于是直接在伯尼·桑德斯的极左激进宣言上签字画押。
你是在亲手给极端的民粹主义递刀子。

硅谷这帮人有一种傲慢的错觉。
他们觉得可以随意白嫖时代的红利，觉得玩火不会烧到自己的手。
但这早已经不是那个世界了。

普通美国人已经明确嗅到了危险。
这股由技术焦虑和失业恐慌点燃的怒火，极其易燃易爆。
Karp 的原话是：“这场火最终会吞噬我们所有人。”

科技确实在颠覆一切。
但如果毫无顾忌地击穿社会底线，资本和技术的反噬，一定会比 AI 的进化来得更猛烈。

[Jielin Dong 董洁林 @D931931]

非常值得关注的是传说中的AI- Native companies，它们的产品服务和技术、公司构架都可以从零开始，它们将抢现有公司的蛋糕、还是创造新需求？

[Jielin Dong 董洁林 @D931931]

AI造成的企业转型过程比想象的要复杂很多。现有企业目前还看不见AI提升其营销规模的迹象，但投入AI的成本是立即的，那么这个预算来自哪里？
1，用公司利润投入AI，这样财务数据立刻受影响，股东不高兴，管理层有压力；
2，如果AI提升了员工效率，可以减少员工，从而腾出预算投入AI，但这个过程比想象的要漫长，各公司基本上处于调研阶段，现在只是停止招新人，而传说中的由于AI而大幅解雇人的过程还没有开始；
3，压缩传统SaaS软件预算、腾出来投入AI，这个过程正在进行中，但这也是个渐进过程，不可能很快释放大笔资金，况且这笔资金的天花板也许比AI所需要小很多。

[Jielin Dong 董洁林 @D931931]

Thank you

[Daniel Lurie 丹尼爾·羅偉 @DanielLurie]

17 months ago, San Francisco stopped spending more than we can afford. The budget I proposed this week, reduces our long-term structural deficit by approximately $300 million dollars.

When I came into office, City Hall was using one-time funds to cover a growing deficit and leaving the bill for someone else to pay.

That wasn’t sustainable. And it wasn’t fair to the taxpayers and the people who depend on city services.
Under my administration, we will not fall back on temporary fixes that force deeper cuts year after year.

We didn’t tell departments to do more with less. We said: focus on what matters most to our recovery and deliver results.

[Jielin Dong 董洁林 @D931931]

37年了，time flies…那一天，我们一群人在Argonne国家实验室公园野炊，有人带了一个电视机，当时CNN在实时播放来自北京的消息，突然电视画面黑了，记者的声音颤抖着说：开枪了、开枪了……

[Vivian Wu @vivianwubeijing]

六四，并不只属于少数的学生抗议者和死难者。所有在北京，在中国，在世界每个角落，经历过听闻过八九年六月三日深夜至六月四日凌晨军队镇压和屠杀的中国人，无论你们手持长枪还是手无寸铁，无论你们是在街头在广场在家中在旅途中，无论你们是否听到过枪声坦克声，所有人其实都经过六四。只不过有的人用光荣和坚持继续走，有些人苟且偷生忘记并转而作恶，有些人假装不知道，有一些人真的不知道也不想知道，然后他们的后代也继续不知道和假装不知道。但六四是中国政治生活公共生活绕不过去也不应该绕过去的坎儿。每个人，都应该有这个认识。
https://t.co/4TdlzL42Mb

[Jielin Dong 董洁林 @D931931]

This is great news—San Francisco could use more businesses, since downtown still has a lot of empty offices and commercial space sitting idle.

[Garry Tan @garrytan]

Prop D was defeated even though the hard leftists spent more than $3M to try to stick it to businesses and working and middle class citizens by dropping an 800% gross receipts tax

Voters saw through the fake “overpaid CEO” marketing lies