The First Batch of Big Tech Employees Laid Off by AI Have Already Returned to Work

Original | Odaily Planet Daily (@OdailyChina)

Author | Golem (@web 3_golem)

The first group of employees laid off by AI has already returned to work.

On February 27, Jack Dorsey (founder of Twitter)’s fintech company Block laid off over 4,000 employees at once, reducing the total staff from 10,000 to less than 6,000. Jack explained the layoffs with “AI tools have changed everything.” While it’s widely accepted that AI will eventually eliminate some jobs, the initial focus on replacing white-collar jobs at middle and high levels has only intensified workplace anxiety. (Related reading: Jack Dorsey’s company, 4,000 white-collar workers being replaced by AI)

However, less than a month later, some of the laid-off employees have already received rehire invitations…

According to Business Insider, these re-employed staff come from various departments, including engineering and recruitment. A design engineer at Block posted on LinkedIn that a leader told him he was laid off by mistake, due to a “clerical error”; an HR person, in a deleted post, said she was re-hired after her manager kept advocating for her; and others said they were mysteriously called back by Block a week after being laid off.

Jack has not publicly responded to the rehire situation. The proportion of re-employed staff is small compared to the total laid off, but it already indicates a problem: some roles and tasks are simply not replaceable by AI.

From a cost perspective, employing an AI at the enterprise level is definitely more expensive than hiring a human.

It costs money to hire people, and tokens to run AI. Claude Opus4.6’s standard base price is $5 per 1 million tokens for input and $25 per 1 million tokens for output; domestic large models are cheaper, with Qwen3.5 plus costing 0.8 yuan per 1 million tokens for input and 4.8 yuan for output.

Take the popular OpenClaw as an example. An internal veteran “shrimp farmer” at Odaily Planet Daily said he used OpenClaw as a life and research assistant for over a month, burning through about $6,000 worth of tokens (using Claude 4.5/4.6 models). $6,000 a month—what kind of high-level intellectual can’t be hired (except in Europe and America)?

If personal use already costs this much, the cost of integrating AI into enterprise work is even higher. For example, replacing customer service with AI— in some regions with educational inflation—spending 3,000 yuan can hire a handsome college student as a customer service rep. But training an AI that can truly replace human customer service, handle complex tickets, connect multiple knowledge bases, conduct multi-turn conversations, and operate stably, definitely costs more than 3,000 yuan a month.

In 2024, Swedish payment company Klarna announced a high-profile layoff of over 1,000 employees, claiming AI customer service could replace the workload of 700 customer service agents. But by May 2025, multiple media outlets including Bloomberg reported that Klarna had started rehiring customer service staff, with the CEO admitting they had “moved too fast” on AI.

Additionally, AI replacing human labor also involves the “Jevons Paradox.”

The Jevons Paradox is an economic concept stating that efficiency improvements don’t necessarily reduce resource consumption—in fact, they can lead to increased total usage due to lower costs and expanded demand. Translated into the AI era workplace, as AI advances improve employee productivity, companies won’t allow employees to rest; instead, they’ll demand more work within the same time frame.

What’s called efficiency actually becomes a more covert form of burden, and AI’s promise of liberating human labor is a complete illusion.

Capitalists also believe that in the AI era, companies need fewer employees, as Jack said, “smaller teams with more intelligent tools.” But in reality? After layoffs, companies aren’t simply replacing jobs with AI; remaining employees are taking on more work with AI’s help.

If it’s just about individual tasks, that’s one thing. But fundamentally, a company is a human organization—where there are organizations, there are “cliques.” AI can integrate into formal organizational structures but will never understand or embed into informal or hidden organizational networks.

When layoffs happen, it’s not just labor that’s cut; organizational “muscle” is also lost. Remaining staff not only shoulder increased workloads but also absorb the anxiety, risks, and responsibilities of their former roles. Fewer collaborators, fewer executors, and most importantly, fewer scapegoats.

During Nvidia’s GTC 2026, CEO Jensen Huang criticized companies that use AI-driven layoffs: “Leaders who rely on layoffs to cope with AI are just out of ideas, their minds are already empty. Even with powerful tools, they won’t use them to expand.” That’s Huang’s direct quote.

Huang’s point is that AI isn’t meant to eliminate employees but to help companies expand and develop new businesses. Instead of layoffs, they should hire more. If management doesn’t realize this, they’re fools. But joking aside, most corporate managers are highly intelligent—they know well the high costs of AI and the ongoing necessity of human labor.

Tech companies’ layoffs may be just a cover; cost reduction is the real goal.

AI has become the universal excuse for tech layoffs. In reality, AI isn’t eliminating individuals but is targeting outdated companies and businesses. When a company can’t keep up with AI advancements, leading to stagnating growth and shrinking profits, the AI revolution becomes a new tool for PUA—reducing staff, cutting costs, and pushing more work onto remaining employees. Then, everyone is left to reflect: why didn’t you become someone better suited for the AI era?

If you’re unlucky enough to hit a “major artery,” you can always quietly rehire later. This approach is common in Silicon Valley. After Elon Musk acquired Twitter in October 2022, he laid off about half the staff (over 3,000 people) in early November, then rehired dozens of those laid off after realizing some key roles couldn’t be left vacant or after discovering mis-hires.

Back to the present: AI will change many things, but it’s not yet magical enough to compensate for strategic blindness, business aging, or lazy management. The story of layoffs and rehires driven by AI, whether due to companies realizing some jobs won’t disappear just because “AI changed everything,” or simply cost-cutting, isn’t inspiring or revolutionary.

It only shows that before the future truly arrives, some have already been hurt by it in advance.