A Thought-Provoking Long Read on Work in the AI Era: Will Weekly Meetings, KPIs, and the 8-Hour Workday Collapse?

Certainly! Below is the English translation of your provided Chinese text. The HTML structure and inline styles have been preserved as closely as possible, while all textual content has been accurately translated into fluent, natural-sounding English suitable for publication. ```html

2026-01-02 09:06

A Thought-Provoking Long Read on Work in the AI Era: Will Weekly Meetings, KPIs, and the 8-Hour Workday Collapse?

Intellectuals zsfzFollow

At the close of 2025, Ivan Zhao (Zhao Yi), founder and CEO of Notion—a U.S.-based unicorn startup valued at billions—published a long-form essay titled “Steam, Steel, and Infinite Minds.”

In this deeply insightful piece, he opens up a grand perspective:

AI is not merely an upgrade to software—it is the “foundational material” of our era. Just as steam engines and steel powered the Industrial Revolution a century ago, AI is now reshaping the fundamental logic of civilization itself.

As an AI entrepreneur focused on enhancing workplace productivity, Ivan paints a breathtaking vision of the future:

Enterprises will be driven by “infinite minds”—never-sleeping intelligences. In that world, countless AI agents will take over all repetitive tasks—

from meeting minutes and knowledge synthesis to handling complex IT requests and writing weekly reports. Companies will no longer be mere aggregations of human labor but symbiotic organizations deeply integrated with AI agents and humans alike.

Lengthy multi-hour meetings will shrink to 10-minute asynchronous reviews.

People will work from anywhere in the world, workflows will run 24/7 across time zones without waiting for anyone to wake up, and decisions will be made through “just-right human-AI collaboration.”

By then, the social rhythms we know today will be utterly disrupted:

“Weekly stand-ups, quarterly planning, annual performance reviews—these industrial-era cadences may soon become obsolete;

they will likely give way to entirely new rhythms yet to be named. We may lose some familiar structures—but gain unprecedented scale and speed in return.”

Yet Ivan Zhao also cautions that on the path toward this future, we face deep-seated “cognitive inertia.”

He quotes Marshall McLuhan’s famous line: “We look at the future through the rearview mirror.”

In the early stages of technological change, people habitually force new phenomena into old frameworks. Ivan sharply observes:

Compared to programmers who have already evolved into “fully autonomous mode,” most knowledge workers today are still frantically pedaling bicycles on the information superhighway.

Humans act like “glue,” exhausting themselves shuttling between countless browser tabs and fragmented pieces of information.

Even our pride in “human-in-the-loop” (manual intervention) can sometimes become a mental shackle.

When automobiles were first invented, laws actually required a person to walk ahead waving a red flag—an absurd constraint. Similarly, when we insist on inserting ourselves into every execution step, we may be inadvertently stifling the full power of the “infinite mind” engine.

This essay is not just about tool evolution—it is a manifesto about the future of organizations, economies, and even human purpose.

It invites us to ponder: when “mind” becomes an infinitely abundant resource, how should humanity redefine itself?

Standing at the dawn of 2026, whether or not we still gaze into the future through the rearview mirror, a radically different era is already roaring toward us, crashing against the edges of the old world.

At this threshold of transformation, Intellectuals presents a full translation and recommendation of this essay, hoping to jointly explore with readers the human-AI symbiosis blueprint behind “Steam, Steel, and Infinite Minds.”

This article originally appeared on the WeChat public account:IntellectualsTranslated by: Meibao. Featured image generated by AI.

Every era has its “miracle material.” Steel forged the Gilded Age; semiconductors launched the digital age. Now, artificial intelligence arrives as “infinite minds.” History teaches us one truth:Those who master the material define the era.

In the 1850s, a young Andrew Carnegie ran through muddy streets in Pittsburgh delivering telegrams. At that time, 60% of Americans were farmers.

Within just two generations, Carnegie and his peers “forged” the modern world. Carriages gave way to railroads, candlelight was replaced by electricity, and iron evolved into steel.

Since then, labor shifted from factories to offices. Today, I run a software company in San Francisco, building tools for millions of knowledge workers.

Here, everyone talks about AGI (Artificial General Intelligence). Yet the world’s two billion office workers have largely yet to feel its real presence.

What will the future of knowledge work look like?

What happens when real-world organizations absorb never-sleeping minds?

The future is hard to predict because we often fail to recognize it when it arrives. Early telephones mimicked telegrams in brevity; early films resembled recorded stage plays.

As Marshall McLuhan once said:

“We look at the future through the rearview mirror.”Today, AI most commonly appears as a Google-like search box. We are stuck in the awkward transitional phase that accompanies every major technological leap.

I cannot foresee everything the future holds—but I like to use historical metaphors to imagine how AI might operate at the levels of individuals, organizations, and entire economies.

I. The Individual: From Bicycle to Automobile

The earliest signs of change have emerged among elite knowledge workers—programmers.

My co-founder Simon was once known as a “10x programmer,” yet now rarely writes code himself.

Walking past his desk, you’d see him orchestrating three or four AI coding agents simultaneously—agents that type faster and think deeper, collectively making Simon a 30x–40x more effective engineer.

He assigns tasks during lunch or before bed and lets them keep working while he’s away. He has become a conductor of “infinite minds.”

In the 1980s, Steve Jobs likened the personal computer to a “bicycle for the mind.” Ever since, we’ve been cycling on the information superhighway for decades.

Yet most knowledge work remains human-powered—like pedaling a bike on a highway. With AI agents, Simon has transitioned from cyclist to driver.

Why is AI assistance harder to implement in general knowledge work than in programming? Because knowledge work is often contextually fragmented, information is scattered, and outputs—unlike code—can’t be easily validated through tests and errors.

When will other knowledge workers get their “cars”? Two problems must be solved:

First,context fragmentation.In programming, tools and context are usually centralized:

IDEs, code repositories, terminals. But general knowledge work is scattered across dozens of tools.

Imagine an AI agent drafting a product brief—it would need to pull insights from Slack threads, strategy docs, last quarter’s dashboard metrics, and even memories stored only in human minds.

Today, humans act as glue—copy-pasting and switching browser tabs to stitch these fragments together.

Unless context is unified, AI agents will remain trapped in narrow use cases, unable to scale.

Second,lack of verifiability.Code has a magical property: it can be verified through tests and error correction.

Model developers leverage this to train AI to improve its coding ability.

But how do we verify whether project management is “good”? We haven’t found a reliable method yet—so we still rely on

(“human-in-the-loop”—a paradigm in AI where human oversight, review, or intervention is required in system decisions or execution)to supervise, guide, and demonstrate what “good” looks like.

The UK’s 1865 Red Flag Act mandated that a person carrying a red flag must walk ahead of any self-propelled vehicle on public roads(repealed in 1896)—a classic example of “human-in-the-loop” backfiring.

Today’s programming agents show us:“Human-in-the-loop” isn’t always wise—just as inspecting every bolt on an assembly line, or requiring a flag-bearer as in the 1865 law, only slows progress. Humans should oversee from a high-level loop, not get stuck inside it.

Once context is unified and work becomes verifiable, billions of knowledge workers will shift from pedaling to driving—and eventually to autonomous operation.

II. Organizations: Steel and Steam

The modern corporation is a recent invention—one that degrades and hits limits as it scales. Centuries ago, most companies were small workshops of a dozen people;

today, we have multinational enterprises with hundreds of thousands of employees. Human brains connected by meetings and messages buckle under exponential communication loads.

We try to mitigate this with hierarchies, processes, and documentation—but that’s like building skyscrapers with wood, using hand tools to solve industrial-scale problems.

Two historical metaphors reveal how new “miracle materials” will reshape future organizational structures.

The first is steel. In the 19th century, buildings topped out at six or seven stories. Iron was strong but brittle and heavy—adding floors risked collapse.

Steel changed everything: strong yet flexible, enabling lighter frames and thinner walls. Skyscrapers soared, and new architectural forms emerged.

AI for organizations is like steel: it canmaintain contextual continuity across workflows and surface decisions while filtering noise when needed.

Human communication no longer needs to serve as load-bearing walls. Two-hour weekly syncs can shrink to five-minute asynchronous reviews;

decisions that once required three layers of approval could be finalized in minutes. Companies can truly scale without the decay previously deemed inevitable.

The second metaphor is the steam engine. Early textile mills were built beside rivers, powered by water wheels.

When steam engines first arrived, mill owners simply replaced water wheels but kept everything else unchanged—productivity barely improved.

The real breakthrough came only when they abandoned rivers altogether.

They built larger factories near workers, ports, and raw materials—and redesigned layouts around steam power. Productivity exploded, launching the Second Industrial Revolution.

This is a waterwheel-powered mill. Though powerful, water flow is inconsistent—limiting mill locations and making them vulnerable to seasonal changes.

We are still in the “replace the waterwheel” phase. AI chatbots are bolted onto legacy tools without reimagining organizational design.

Once old constraints dissolve and companies are powered by never-sleeping minds, I can envision entirely new possibilities.

At Notion, we’ve already experimented: beyond our 1,000 human employees, over 700 AI agents handle repetitive tasks.

They take meeting notes, answer questions, synthesize knowledge, process IT requests, archive customer feedback, onboard new hires, write weekly reports, and more.

This is just the beginning.The true upside is limited only by imagination and inertia.

III. The Economy: From Florence to Megalopolis

Steel and steam transformed not just buildings and factories—but cities. Centuries ago, cities were built on a human scale—you could cross Florence in 40 minutes.

Steel enabled skyscrapers; steam connected urban cores to hinterlands; elevators, subways, and highways followed.

Cities exploded in scale and density: Tokyo, Chongqing, Dallas… These aren’t just enlarged Florences—they represent entirely new ways of living.

Megalopolises can feel disorienting and overwhelming—yet they offer more opportunity and freedom.

More people do more things in more diverse ways—richness and complexity far exceeding what a walkable, intimate Renaissance city could ever contain.

Today, knowledge work accounts for nearly half of U.S. GDP. Yet its operational model remains largely human-scaled: teams of dozens, processes driven by meetings and emails, organizations that strain under modest headcounts.We’ve built Florences out of stone and timber.

When AI agents deploy at scale, we’ll start building Tokyos—massive organizations spanning thousands of AI agents and humans;

workflows running 24/7 across time zones, never waiting for anyone to wake up;decisions made through “just-right human-AI collaboration.”

Living within such systems will feel profoundly different: faster, less effortful—but initially dizzying.

The rhythms of weekly meetings, quarterly planning, and annual reviews may no longer fit; new cadences will quietly emerge.

We may lose some familiar order—but gain unprecedented scale and speed in return.

Every miracle material has forced humanity to stop looking through the rearview mirror and start imagining entirely new worlds.

We’re still in AI’s “waterwheel phase”—awkwardly bolting chatbots onto existing workflows.

We must stop treating AI as just a co-pilotand instead ask: what will knowledge work look like when mundane tasks are entrusted to never-sleeping minds?

Steel. Steam. Infinite Minds. The next skyline is already visible—we need only build it ourselves.

This article originally appeared on the WeChat public account:Intellectuals

AI Venture DailyChannel: Frontier Tech

Published with the author's permission. Views expressed are solely the author’s and do not represent Huxiu’s position.
For feedback or complaints regarding this article, please contact tougao@huxiu.com.

Those changing the world—and those who want to—are all on Huxiu App

``` > **Note**:    > - Font names like `'微软雅黑'` and `'宋体'` have been converted to standard English equivalents (`'Microsoft YaHei'` and `'SimSun'`) for better cross-platform compatibility.    > - All technical terms (e.g., “human-in-the-loop,” “AGI”) are preserved with explanations where originally provided.    > - The tone balances journalistic clarity with philosophical depth, matching the original’s blend of futurism and historical analogy. Let me know if you'd like a plain-text version or further formatting adjustments!

2026-01-02 09:06

震撼长文解读AI 时代的工作 ：周会、KPI、8小时工作制会崩溃吗？

知识分子zsfz关注

2025 年岁末，美国百亿独角兽公司Notion 创始人兼 CEO Ivan Zhao（赵伊）发表了长文《蒸汽、钢铁与无限心智》

（Steam， Steel， and Infinite Minds）。

在这篇**洞见性的文章中，他开启了一个宏大的视角：

AI 不仅仅是软件的升级，而是这个时代的“基础材料”——正如同百年前的蒸汽机与钢铁之于工业革命，AI 正在重塑文明的底层逻辑。

作为一名专注于如何提升办公效率的AI创业者，Ivan 描绘了一幅令人震撼的未来图景：

企业将由“永不休眠的心智”（Infinite Minds）驱动。在那个世界里，无数 AI Agent（智能体）承揽了所有的重复性劳动——

从会议记录、知识整合到处理复杂的 IT 请求和撰写周报。公司不再是单纯的人力集合，而是一个由 AI 代理与人类深度耦合的共生组织。

曾经冗长的几小时的会议将缩减为10分钟的异步审阅，

人们可以分布在世界的任何地方工作，工作流昼夜不息，跨时区推进，不用等待任何人醒来，决策都在“恰到好处的人机协同”中完成。

到那时，我们熟悉的社会节律将被彻底打破：

“周会、季度规划、年度考核……这些工业时代的刻度或许将不再合时宜；

取而代之的，是某种全新的、尚未被命名的节律。我们或许会失去一些熟悉的秩序，却将换来前所未有的规模与速度。”

然而，Ivan Zhao同样提到，在通往未来的路上，我们仍面临着深刻的“认知惯性”。

借用麦克卢汉那句名言：“我们总是看着后视镜驶向未来。”

在技术变革的初期，人们习惯于用旧的框架去硬套新的事物。Ivan 尖锐地指出：

相比于已经进化到“全自动模式”的程序员，大多数知识工作者的现状仍像是在高速公路上拼命蹬自行车。

人类像“胶水”一样，在无数标签页和碎片化信息间疲于奔命。

甚至，我们引以为傲的“Human in the loop”（人工介入）有时也是一种思维枷锁。

正如汽车刚发明时，法律竟要求派人举着红旗在车前开道——当我们在追求深度介入每一个执行环节时，或许正是在阻碍那台“无限心智”引擎的发力。

这篇文章不仅是关于工具的进化，更是一份关于组织、经济乃至人类存在价值的未来宣言。

它邀请我们思考：当“心智”成为无限供给的资源，人类该如何重新定义自己？

站在 2026 年的破晓时刻。无论我们是否仍习惯于“透过后视镜观察未来”，一个截然不同的时代已然呼啸而至，撞击着旧日的边缘。

在这个万象启新的起点，《知识分子》特全文编译并推介了此文，希望与读者们一起共同探讨《蒸汽、钢铁与无限心智》背后的人机共生蓝图。

本文来自微信公众号：知识分子，编译：梅宝，题图来自：AI生成

每个时代都有它的“奇迹材料”。钢铁铸就了镀金年代，半导体开启了数字时代。如今，人工智能以“无限心智”之姿降临。历史告诉我们：掌握材料者，定义时代。

在19世纪50年代的匹兹堡，少年安德鲁·卡内基奔跑在泥泞的街巷，做着电报员的工作。当时六成美国人是农民。

仅仅两代人之后，卡内基与他的同辈们便“锻造”出了现代世界。马车让位于铁路，烛光被电力取代，铁器升级为钢材。

自此，劳动从工厂转入办公室。今天，我在旧金山经营一家软件公司，为数百万知识工作者打造工具。

在这里，人人都在谈论AGI（Artificial General Intelligence ，通用人工智能）。但全球二十亿办公族大多尚未真切感受它的存在。

知识工作的未来会是怎样的？

当现实社会中的组织结构吸纳了永不休眠的心智，又会发生什么？

未来常难以预测，因为我们常常识别不出来。早期的电话像电报般简短，早期电影宛如舞台剧的录像。

正如麦克卢汉（Marshall McLuhan）所言：

“我们总是透过后视镜驶向未来”。如今的AI，最常见的形态仍像昔日的谷歌搜索框。我们正处在每一次技术跃迁都会经历的尴尬过渡期。

未来会如何，我并不能尽知。但我喜欢用几则历史隐喻，去推想 AI 在个体、组织乃至整体经济等不同尺度上的运作方式。

一、个体：从自行车到汽车

最早的变化迹象，出现在知识工作的顶尖群体——程序员中。

我的联合创始人西蒙曾是所谓的“十倍速程序员”，如今却很少亲自敲代码。

走过他的桌前，你会看见他同时调度着三四个AI编程代理，它们不仅打字更快，还更能思考，合起来让西蒙成为三十至四十倍效率的工程师。

他在午休或睡前布置任务，任由它们在自己离开时持续工作。他成了“无限心智”的指挥者。

上世纪80年代，乔布斯（Steve Jobs）把个人电脑比喻成“头脑自行车”。自那以后，我们在信息高速公路上骑行了数十年。

然而，多数知识工作至今仍由人力驱动，就像在高速公路上骑单车。有了AI代理，西蒙已从骑行者变为驾驶者。

相较于编程代理，AI 为何在辅助知识工作上更为困难？原因在于，知识工作往往情境分散、信息零碎，且成果难以像代码那样通过测试与错误来直接验证。

其他知识工作者何时能换上“汽车”？须解决两个问题：

其一，情境碎片化。在编程领域，工具与上下文往往集中在同一处：

集成开发环境（IDE）、代码仓库、终端。而一般的知识工作却分散在几十种工具之间。

试想，一个 AI 代理要起草一份产品简报，它需要从 Slack 的讨论串、战略文档、仪表盘里的上季度指标，以及仅存在于人脑中的记忆中提取信息。

如今，人类就像胶水，用复制粘贴与切换浏览器标签页等方式，把这些碎片粘合在一起。

除非能将情境整合归一，否则 AI 代理只能困在狭窄的应用场景里，难以拓展。

其二，可验证性缺失。代码有个奇妙特性：你可以用测试与错误来进行验证。

模型制作者借此训练AI提升其编程能力。

但如何验证项目管理是否得当？我们尚未找到合适的方法，因此仍需人在回路

（human in the loop，指的是在人工智能领域，系统的决策或执行过程中，必须有人类参与、审核或干预的一种工作模式），进行监督、引导并示范何为“好”。

1865 年的《红旗法案》（The Red Flag Act）规定，车辆行驶在街上时，必须有一名持旗者在车前步行开路（该法案于 1896 年废除）。这是“人在回路”适得其反的典型例子。

今天的编程代理则启示我们：“人在回路”未必可取——犹如在流水线逐一检视每颗螺栓，或如1865年《红旗法案》要求有人持旗步行在前清道。我们应让人类在高位监督回路，而非陷于其中。

一旦情境整合、工作可验证，亿万知识劳动者将从骑行迈向驾驶，继而进入自动驾驶。

二、组织：钢铁与蒸汽

公司是近代发明，会随规模扩张而衰减并触顶。几百年前，多数公司是十几人的作坊；

如今，我们有数十万人的跨国企业。以会议与消息相连的人脑沟通，在指数级负荷下不堪重负。

我们尝试用层级、流程与文档来缓解，但这如同用木材建造摩天楼，以人力工具解决工业级难题。

两个历史隐喻昭示新的奇迹材料如何重塑未来的组织架构。

**个是钢铁。19世纪，建筑限高六七层。铁虽强却脆且重，加层会塌。

直到钢材改变了一切，它强而韧，框架更轻、墙体更薄。楼宇可拔地数十层，新建筑形态应运而生。

AI之于组织，恰似钢铁：它有潜力在工作流间保持情境连贯，并在需要时呈现决策、屏蔽噪音。

人类沟通不必再充当承重墙。每周两小时的同步会可缩短为五分钟的异步审阅；

需经过三级审批的高管决策或可在数分钟内完成。公司得以真正规模化，而不必承受以往被视为必然的衰减。

第二个是蒸汽机。工业革命初期，纺织厂傍河而建，靠水轮驱动。

蒸汽机问世之初，厂主仅换掉水轮，其余都照旧，生产力并没有大的提升。

真正的突破来自彻底摆脱河流。

在靠近工人、港口与原料处建更大厂房，并围绕蒸汽机重新设计布局。生产力爆发，第二次工业革命腾飞。

这是一座以水轮为动力运转的磨坊。水力虽强劲，却不稳定，这使磨坊只能设在有限的地点，并受季节变化制约。

我们仍处于“换掉水轮”的阶段。AI聊天机器人外挂于原有工具，尚未重构组织形态。

当旧约束消解，公司可由永不休眠的心智驱动，我便能想象全新的可能。

在我所在的Notion公司，我们已经试验过：在千名员工之外，七百余个AI代理承担了重复工作。

它们记录会议、答疑整合知识、处理IT请求、归档客户反馈、协助新人入职、撰写周报等。

这仅是起步，真正的收益只受想象力与惰性的限制。

三、经济：从佛罗伦萨到超级都市

钢铁与蒸汽改变的不仅是楼宇与工厂，还有城市。几百年前，城市依人而建——佛罗伦萨四十分钟便可穿城。

钢材使摩天楼成为可能，蒸汽机连通了城市中心与腹地，电梯、地铁、公路接踵而至。

城市在规模与密度上爆发：东京、重庆、达拉斯……它们不只是放大的佛罗伦萨，而是全新的生存方式。

超级都市令人迷失、难行，但也带来了更多机遇、自由。

越来越多的人以越来越多样的方式，做越来越多的事——其丰富与复杂度，早已超出一个人行可达、人际亲密的文艺复兴古城所能容纳的极限。

如今，知识工作已占美国 GDP 近半。但其运行方式多半仍停留在人力尺度：数十人的团队、由会议与邮件牵动的流程、人数一多便捉襟见肘的组织。我们就像用石头与木材，筑起了一座座佛罗伦萨。

而当 AI 代理大规模上线，我们将转而营建东京——横跨成千上万AI代理与人类的巨型组织；

工作流昼夜不息、跨时区推进，不再等待任何人醒来；决策在“恰到好处的人机协同”中完成。

置身其中，感受会截然不同：更快、更省力，但起初也会令人晕眩。

周会、季度规划、年度考核的节拍，或许不再合乎时宜；新的节律会悄然成型。

我们或许失去了一些熟悉的秩序，却换来了前所未有的规模与速度。

每一种奇迹材料，都曾逼使人们不再透过后视镜审视世界，而开始想象全新的图景。

我们仍处在 AI 的“水轮阶段”——把聊天机器人生硬地加装到工作流上。

我们必须停止只让 AI 做副驾，而要畅想：当繁杂琐事被托付给永不休眠的心智，知识工作将呈现怎样的面貌？

钢铁，蒸汽，无限心智。下一座天际线已在眼前，只待我们亲手筑起。

本文来自微信公众号：知识分子

AI创投日报频道: 前沿科技

本内容由作者授权发布，观点仅代表作者本人，不代表虎嗅立场。
如对本稿件有异议或投诉，请联系 tougao@huxiu.com。

正在改变与想要改变世界的人，都在 虎嗅APP