组织权力的起源：从地理跨度到信息路由
在过去近两百年里，大型企业面临的核心挑战是如何协调地理上分散、无法直接沟通的成千上万名员工。中层管理（Middle Management）应运而生，其核心职业功能是作为企业内部的信息路由枢纽。在19世纪40年代前，美国最大型的企业也仅雇佣几十人，经营者可以亲眼见证整个运作流程。然而，随着铁路系统（Railroad）的兴起，这种小规模经营模式无法适应大规模协调的需求。以1891年的宾夕法尼亚铁路为例，它雇佣了超过11万名员工，处于匹兹堡的列车延误将直接影响巴尔的摩及费城等地的维护与调度人员。在这种环境下，第一层专业管理诞生了：这些人员不具备所有权，其核心使命就是确保各部门人员知晓彼此的动向。当这一行政职能完成标准化后，管理层逐渐发展为具有独立政治属性、利益诉求及自我扩张倾向的机构。
Original English Source
For nearly 200 years, corporations had
the same problem. How do you coordinate
thousands of people who cannot talk to
each other directly? The solution was
middle management. Entire careers were
built around routing information between
people. The problem is AI is starting to
automate the thing that managers
historically did best, the routing of
information. ... The concept of middle management is
approximately 180 years old. It was
invented to solve a very specific
technical problem and that is how to
coordinate work across geographic
distance when there was no other way of
passing the information other than a
horse. Before the 1840s, the largest
business in the United States employed a
few dozen people. A business owner could
see the entire operation with their own
eyes. They had an opportunity to know
every worker by their name because there
was nothing to coordinate at scale until
the Americans began building the
railroad. The Pennsylvania road alone
employed over 110,000 workers by 1891. A
train running late in Pittsburgh had
consequences for workers in
Philadelphia, Freud chippers in
Baltimore, and maintenance crews in
Pittsburgh, none of whom could I know
about each other without an
intermediary. And that intermediate area
would carry the information between the
parties. And this is how the first
professional management layer was born.
A salaried employee with no ownership
stake whose entire job was to coordinate
people and make sure that they're aware
of what others are doing. But once the
admin functions were done, the
management became its own institution
with its own politics, its own
interests, its own career structures,
and its own tendency to self-perpetuate.
从二战扩张到精益生产的批判
二战后，美国企业步入全球化运营阶段，运营版图跨越20多个国家，催生了区域、职能等多层级管理架构，也随之产生了正式的长期规划、预算编制、生产率分析等繁琐的官僚程序。直到20世纪70年代的石油危机、滞胀以及日本制造业（Japanese Manufacturing Competition）的崛起，美国传统管理模式才遭遇了首次严重批判。以丰田（Toyota）为代表的精益生产理念，通过大幅裁撤管理层级、强化一线直接问责制，实现了比美国同类制造业更快、更高效且品质更优的运作。这一模式的成功促使美国公司在随后的20年里致力于组织架构的层级扁平化。最终，这种改革与后期企业的过度扩张，叠加AI时代的降临，使得管理层消除成为当前美国企业界裁员的最核心目标。
Original English Source
When the original problem got solved,
the post-war period created conditions
for the expansion of middle management.
American companies became global after
World War II. There were companies that
had operations in more than 20 countries
that required more layers of regional,
divisional, and functional managers. And
all of that gave the rise to the concept
of formal long range planning. The
5-year plans, the budgeting, reports,
summaries of completed work, analysis of
productivity, and bureaucracy. The 1973
oil crisis, stagflation, and the
emergence of Japanese manufacturing
competition in the late '7s produced the
first serious critique of American metal
management. The Japanese companies and
particularly Toyota came up with a
concept of lean production that cut
management layers hard and became much
faster and more productive than the
American manufacturing without losing
quality. American companies looked at
that, saw fewer layers, direct
accountability from workers, workers
empowered to stop production lines and
solve problems without asking a
supervisor and said, "We're going to do
the same." And so for the next 20 years,
it was a phase of delayering,
relayering, and all kinds of management
permutations until we got into CO and
CO's infamous overhiring that then
collided with the expansion of AI and
the mass layoffs that followed made
management elimination the most targeted
layoff category in corporate America
right now.
管理工具的实践优化
(本视频作者在此处分享了其关于如何通过 AI 构建新闻采集 workflow 的实践，该 workflow 依托于 Hostinger 的 VPS 环境，自动化了信息筛选与 Notion 记录流程，旨在替代繁琐的人工调研。)
Original English Source
Let me show you a bit of the behind the
scenes of how we run this channel. When
I'm preparing for videos, I do a lot of
research. And to make my life a little
bit easier, I have a news collector
workflow that was built using OpenClaw
on Hostinger virtual private server. And
this workflow automates part of my
research process. Every day, it scans AI
and tech news sources, filters out the
noise, collects only the important news,
applies tags, and adds everything into
my notion database. It also sends me a
telegram message and notifies when a new
source has been added. ... (Detailed promotion of Hostinger's setup for AI agents)
裁员的金融逻辑：效率至上
以客观视角审视，企业在制定财务模型时，每一位雇员都必须带来可量化的财务成果，否则将被更经济的替代品所取代。裁减中层管理存在明确的财务核算：以某大型科技公司为例，通过裁撤大量管理岗位，一年可节省约30亿美元。市场对此类行动的反应通常积极，往往伴随着股价上涨。现阶段，企业进行裁员决策的主要衡量标准是单位员工产出营收（Revenue per Employee）。虽然营收增长与裁员缩减 headcount 两者往往交织，难以完全剥离具体影响因素，但这一指标的变化确实支持了市场偏好扁平化管理架构的观点。
Original English Source
30% of all layoffs in 2024 was middle
management. US middle management job
postings are 42% lower than the previous
years. Now, if you take the human aspect
out of the equation, when a business or
a founder or an entrepreneur calculates
the financial model for their business,
every employee has to bring a measurable
financial result or if the decision to
hire somebody is going to influence an
actual financial result. If an employee
does not bring a measurable result, you
either get rid of that employee or you
replace it with something cheaper. So,
in the interest of being objective,
let's do the math. The math of cutting
middle management. ... (Morgan Stanley AWS calculation example) ... The corporate lingo for fewer
managers. ... Following the announcement, Block's
stock rose 25% in a single day. The
market reaction to middle management
cuts has been consistently positive in
the short term. ... As of early
2026, the leading metric companies use
to make decision about cuts is revenue
per employee. It is the most tracked
metric across current restructurings and
a good one to use to measure whether
layoffs are actually making a
difference. And as you can see, Meta's
revenue per employee almost doubled.
Blocks quadrupled. Amazon numbers are
still unknown but trending up.
组织结构重构：从层级到智能
2026年初发布的《从层级到智能》（From Hierarchy to Intelligence）一文指出，过去2000年来，组织层级的功能仅是作为信息路由协议。由于单个领导者有效管理的跨度有限（通常为8人），因此在传递信息时必须构建层级。该文的核心论点是：AI完全能够取代这一信息路由功能。在此背景下，一种新的三角色组织结构被提出：
- 直接接收模型上下文的贡献者（Individual Contributors）
- 负责跨部门问题的关键责任人（Directly Responsible Individuals）
- 边建设边赋能的教练型角色（Player-Coach Profile）
这种结构认为，在后AI时代，永久性的中层管理层已无存在的必要。
Original English Source
On March 31st, 2026, exactly 5
weeks after Block's layoff announcement,
Jack Dorsey and Rule of Botha published
an essay titled From Hierarchy to
Intelligence and released it on Block's
website and Sequia Capital's platform.
The essay attracted 5 million views in
the first 48 hours. The core argument of
the essay is that for 2,000 years from
the Roman Empire to the Prussian
military to American railroads and
corporations, the organizational
hierarchy functioned as an information
routing protocol. The logic behind it is
that a single leader could effectively
manage only eight people around them,
which means that to pass the information
to more than eight people, you would
need one person to manage every eight
people. The essay argues that AI can
fully replace this function. In its
place, Block proposes a three role
organizational structure. Individual
contributors who receive context from
the model directly, the directly
responsible individuals who own
crosscutting problems and the player
coach profile who builds alongside their
teams and develops the talent. The
central argument is essentially that
there is no need for a permanent middle
management layer.
经济成本分析：人类经理 vs AI 智能代理
裁撤中层管理后，虽然节省了高额的薪酬开支，但由于采用了 AI 技术，企业会将资金转移至云服务和 AI 资源消耗中。以一个软件开发管理岗为例：一名包含奖金、福利的综合人力成本约为18万美元；而使用托管的 AI 代理平台（如类似 OpenClaw 的方案），基于处理会议纪要、生成更新、 escalation 等复杂自动化场景，即使采用高强度使用配置，其年度成本（包含算力、模型 Token 与工程维护费）也远低于人类岗位的开支。随着单位 Token 价格的下降，对于主要从事协调工作的管理岗而言，这种经济分析天平明显向技术替代倾斜。
Original English Source
... So continuing
with the calculation, what happens when
you eliminate the middle management
layer and automate coordination work
that they used to do? Because AI apps
that you're going to use to replace the
people aren't cheap, they also cost
quite a bit. What happens is that you
end up with fewer human salaries but
more spent on cloud bills and AI
credits. So what is actually cheaper? A
human manager or agents on a platform
like OpenClaw. Let's do the math. ... (Detailed cost breakdown analysis) ... 18
versus 180. Now, combine this with the
fact that token prices per unit are
falling. ... But nevertheless, if we're
looking at a middle manager without
strong hard skills or someone without
coaching skills that can truly develop
people or someone whose job is primarily
coordination, you can see how the math
is not working in their favor.
未来职业定位：成为承重者而非展示者
在管理岗日益稀薄的环境中，对于处于资深/总监层级 cusp 的人员而言，不能再依赖“假装胜任”来维持职业生涯。未来的管理角色——无论是 Player-Coach 模型还是其他架构——要求管理人员具备：
- 创造可量化价值的能力：例如在美元产出上有直接影响力。
- 深度硬技能与行业知识：掌握复杂的专业流程或行业深度，使其难以被重建。
- 人才培养与领导力：真正的 coaching 能力。
这种趋势下，资深个人贡献者（Senior IC）的平行职业路径变得更具可行性与价值。职业生涯应当致力于成为“承重者”（Loadbearing）——不可或缺、专注实效，而非仅仅追求曝光度或“表面话语”能力。
Original English Source
So what is the middle management role
becoming? The player coach model
actively promoted by block where you're
quite literally a player and a coach. It
is viable if you can genuinely produce
highquality individual contributor work
and develop people effectively. This
model is also objectively much more
demanding because it combines two
completely different skill sets. IC plus
people development for one salary. ... The senior
individual contributor who leads through
expertise and influence rather than
authority. ... There is another trend among Gen Z where
more than half of them are actively
avoiding middle management roles. ... To
sum it up, the middle management layer
is indeed getting thin. And for people
who sit on the cusp of senior/director
level or senior managers, you got to
have a moat. ... In other words, your career as
a middle manager should be oriented
toward becoming loadbearing,
not visible, not liked, not even
well-reed. You have to be loadbearing. ... So to middle managers or
any other role really, it is time to
roll up your sleeves, stay relevant,
stay hungry, stay curious and master
your craft.