Editor’s Brief
A critical look at the 'AI-first' corporate restructuring trend, triggered by Block’s massive layoffs and the subsequent market rally, signaling a fundamental shift in how human capital is valued in the age of cognitive automation.
Key Takeaways
- Block's 40% workforce reduction resulted in a 24% stock price surge, indicating that investors now reward 'de-humanized' efficiency over traditional scale.
- The Klarna case demonstrates that while AI has an 'empathy ceiling,' it has permanently eliminated the need for standardized information-processing roles.
- The 'AI-first' strategy adopted by firms like Duolingo shifts the labor model from 'one person, one job' to 'one person plus AI equals three jobs,' shrinking the total employment pool.
- Unlike previous industrial revolutions that replaced physical labor, AI targets cognitive tasks, leaving white-collar workers with fewer 'higher-ground' roles to retreat to.
Introduction
Fu Sheng shared his observations on Block’s massive layoffs: the payment giant’s stock price soared after cutting nearly half its workforce, revealing a radical shift in corporate logic in the AI era. This is no longer a desperate survival tactic due to poor management, but a proactive reshaping of the organization using AI. As the reality of “one person plus AI equals three” takes hold, the boundaries of traditional knowledge labor are being permanently pushed inward by technology.
Key Takeaways
- Block’s proactive 40% layoff triggered a stock price surge, signaling that companies are shifting from passive response to proactive use of AI for structural efficiency gains.
- The Klarna case proves that while AI has bottlenecks in handling complex emotions, roles involving standardized information integration and inquiries have permanently disappeared.
- The “AI-first” strategy is leading to a reduction in the total number of jobs; the core logic is that one well-educated employee paired with AI can complete several times the workload.
Remarks
Don’t be lulled by the claim that “AI has a ceiling.” Klarna’s fluctuations are merely testing the boundaries of replacement. The real signal is that once-respectable knowledge labor is being rapidly commoditized. For professionals, the key is how to avoid becoming the standardized part that the “AI-first” logic prioritizes for elimination.
“`
* Preserve HTML? Yes.
* No added/removed facts? Yes.
* Numbers/URLs/Names accurate? Yes (40%, 24%, 2026, Fu Sheng, Block, Jack Dorsey).
* Output HTML only? Yes.
(
To be honest, when I saw this news, I felt a bit of panic myself—a young person using a dozen or so agents to replace me, the boss. This isn’t a joke; it’s a real possibility. AI layoffs don’t just affect ordinary white-collar workers. Today, I want to clarify this: what exactly is happening in this round, where the boundaries are, and why this is different from every other technological revolution in history.
On February 26, 2026, Block, the fintech company owned by Twitter founder Jack Dorsey, announced it was laying off 4,000 people—nearly half of its total workforce.
You might not have heard of Block, but you probably know its products: Square, which handles payments for merchants, and Cash App, the American version of “Alipay.” With an annual revenue of $24 billion, it’s not a small company.
On the day of the layoffs, the stock price rose by 24%.
Investors are applauding this. This means—it won’t stop.
"In the coming year, I believe most companies will reach the same conclusion and make similar structural adjustments."
——
igcaption>
These are the publicly acknowledged figures. More layoffs used the term "structural adjustment," without naming AI, but the logic behind them is the same.
Entering 2026, there was no slowdown. In January, Pinterest announced layoffs of nearly 15%, explicitly mentioning AI. Chemical giant Dow Chemical laid off 4,500 people, citing "restructuring all business processes with AI."
The scale is clear. The next question is: where is the boundary—which jobs are disappearing, and which remain?
Klarna’s Experiment: The Boundary is Further In Than You Think
To clarify "where the boundary is," Klarna is the most worthwhile case to look at.
This European fintech company cut about 700 customer service positions between 2022 and 2024, reducing its overall workforce by 40%. The CEO later admitted this was the result of the combined effects of AI and natural attrition.
It started like a textbook—costs down, efficiency up. Then in 2025, the CEO said something else:
"We cut too aggressively."
After replacing customer service with AI on a large scale, customer satisfaction dropped significantly. AI couldn’t handle problems that required understanding complex
…` (Wait, the source ends at “原因只有”).
*Re-checking the end of the source:* `
原因只有`
*Translation:* `
The reason is only`
*Re-checking the beginning:* `t;。但这个结论只对了一半。`
*Translation
一个：以前的技术，替代的是体力。这次替代的是认知本身。
The definition of work used to be exchanging one of your skills for a salary. Whoever could make that board move would get paid. But today, Lobster can make it move by consuming just a few tokens.
Steam engines and robotic arms replaced hands and feet. Those people could move up to do white-collar jobs that required thinking.
But now, what AI is replacing is that very “upward” direction—writing reports, analyzing data, handling customer complaints, creating content, and replying to emails are all cognitive tasks.
Last time: Physical labor was replaced, and people could move up to become white-collar workers.
This time: Cognition is being replaced, and that “upward” direction is also being compressed.
There are still paths toward higher-level cognition, but the capacity of those paths is far smaller than the
Final Thoughts
Mustafa Suleyman, Microsoft’s AI Chief, says that in the next 12 to 18 months, most white-collar tasks can be fully automated by AI. He is a co-founder of DeepMind, specifically recruited by Microsoft to lead their AI strategy. These are the words of an insider.
That 12 to 18-month countdown has already begun.
I am not saying this to make anyone anxious. Anxiety is useless; seeing things clearly is what’s useful.
First, start actually using AI now—don’t just chant slogans. If you don’t use it yourself, you won’t know where the boundaries are, nor will you find your truly irreplaceable position. Only by using it will you know what AI can truly do and where it still falls short.
Second, practice judgment more, don’t just focus on execution. AI is an excellent executor, but it lacks initiative—it won’t proactively identify problems or ask, “Should we be doing this?” Initiative and judgment are currently the hardest things for AI to replace.
Third, reduce fixed debt and leave yourself some breathing room. In the face of structural changes, the ability to resist risk is more important than the ability to attack.
This decade is a period of growing pains, and it’s truly difficult. The next generation will have it better—not because the problems have vanished, but because AI will ultimately create far more than it destroys. However, no one can walk this middle path for you.
The key to making it through is that you must still be there.
Source
Author: Fu Sheng
Published: March 14, 2026, 22:09
Source: Original Post Link
Editorial Comment
The news of Block cutting 4,000 jobs—nearly half its workforce—while its stock price simultaneously jumped 24% is more than just a financial headline. It is a manifesto for a new, colder era of corporate management. As a technology editor, I’ve seen my share of 'downsizing' cycles, but this feels fundamentally different. In the past, massive layoffs were a signal of distress, a desperate attempt to stop the bleeding of a failing enterprise. When Jack Dorsey swings the axe at Block, however, he isn't saving a sinking ship; he is redesigning a functional one to run without the 'friction' of human labor. The market’s rapturous response tells us everything we need to know: in the eyes of capital, human employees are transitioning from assets to be nurtured into liabilities to be optimized out.
Fu Sheng’s observation about the 'young person with a dozen agents' replacing the boss might sound like a Silicon Valley hyperbole, but the underlying logic is terrifyingly sound. For decades, the social contract of the white-collar world was built on the value of 'knowledge work.' We told ourselves that as long as we were educated, as long as we could process information, write reports, and analyze data, we were safe from the reach of automation. We assumed the 'cognitive' was a uniquely human sanctuary. We were wrong. What we are witnessing now is the industrialization of cognition. When intelligence is tokenized and sold as a utility, the 'decent' office job—the kind that involves synthesizing information and producing standardized outputs—becomes as vulnerable as a textile worker in the 19th century.
The Klarna example is particularly instructive for those looking for a silver lining. Many analysts pointed to Klarna’s recent admission that they 'cut too deep' and saw a dip in customer satisfaction as proof that AI has a ceiling. They argue that AI cannot handle complex emotions or nuanced human problems. While true, this is a dangerous comfort. Klarna didn't revert to its original headcount; they simply found the new boundary. The standardized, routine, and predictable tasks are gone forever. They aren't coming back. The 'ceiling' isn't a shield for the majority of workers; it is a narrow ledge that only a few will be able to stand on.
This brings us to the 'AI-first' strategy being touted by companies like Duolingo. This isn't just a buzzword; it’s a structural shift in the volume of labor required to run a global business. If one well-educated worker, armed with a suite of AI agents, can produce the output of three or four people, the math for the middle class simply doesn't add up. We are looking at a future where the 'entry-level' knowledge job disappears. If the AI can do the work of a junior analyst, how does the next generation ever become a senior analyst? We are effectively burning the bottom rungs of the career ladder.
For the individual professional, the takeaway isn't to wait for the storm to pass, because this isn't a storm—it’s a climate shift. The traditional definition of talent is being rewritten in real-time. It is no longer enough to be a 'standardized part' of a corporate machine. If your job can be described in a manual or a set of repeatable steps, an LLM is already being trained to replace you. The only remaining value lies in judgment, initiative, and the ability to ask the questions that AI doesn't know to ask. We are moving from an economy of 'doing' to an economy of 'deciding.'
Ultimately, the Block layoffs signal that the 'AI-first' logic is no longer a pilot program; it is the new baseline for competitiveness. The 12 to 18-month window suggested by industry insiders like Mustafa Suleyman for the automation of white-collar tasks is already ticking. The goal for any professional now shouldn't be to compete with the machine, but to ensure they are the one directing it. Because as the market has clearly shown, it will no longer pay a premium for human effort when silicon can provide the same result for the price of a few tokens.