Is the end of the working class approaching? The wave of layoffs in 2026 has only just begun…

Author: Byron Gilliam

Original Title: Jobpocalypse now?

Translation and Editing: BitpushNews

Even when I worked at the investment bank during its good times, I always felt a new round of layoffs was just around the corner—I think partly because management had no idea how many people they actually needed.

I worked in the sales and trading floor, where every day ended with a revenue figure: client commissions minus trading losses (occasionally profits too). So you might think that quantifying who contributed what, or who caused losses, should be straightforward.

But it’s not.

A commission paid for a trade might be partly or entirely attributable to research analysts, salespeople, or traders who talked to the client—or to the trader on the other side of the trade (that was me!).

No one truly knows why clients choose to trade with us. Therefore, it’s impossible to clearly attribute each commission to a specific individual, making it hard to determine who is absolutely essential to the business.

To borrow (retail giant) WarnerMack’s words, half of the payroll might be wasted; they just don’t know which half.

The only way to find out is to fire some people and see what happens.

It feels like the same thing is about to happen across companies everywhere, because it’s not just investment banks facing this problem.

When work was mainly in agriculture and manufacturing, measuring productivity was easy: just count how many apples they picked or how many parts they produced.

But when most people start working in offices, things get much more complicated.

“Knowledge work isn’t defined by quantity,” Peter Drucker wrote. “Knowledge work isn’t defined by its cost. Knowledge work is defined by its results.”

Employers don’t know how to measure these results—the output units of a day’s meetings, calls, and internal memos?

So they turn to measuring time: employees are asked to spend eight hours in the office each day for pay, and employers expect them to produce eight hours’ worth of work during that time.

Time becomes a proxy for output.

But what happens when everyone works from home?

If employers can’t measure their employees’ time in the office, they have to measure their output instead.

And that’s a good thing. “Focusing on output rather than activity is key to increasing productivity,” Peter Drucker wrote in 1967.

But employers have never really figured out how to do that.

Now, artificial intelligence (AI) is forcing employers to try again. Large language models can handle many time-consuming tasks, prompting companies to rethink what they pay employees to do.

I’m not sure they’ll do better than the banks I worked at. But AI narratives are putting enormous pressure on companies, forcing them to find ways to boost productivity—so much so that many are simply laying off workers to see what happens.

Data from March 6 shows this may already be happening: U.S. Bureau of Labor Statistics reports that last month, employment in the tech sector decreased by 12,000 jobs month-over-month, with a total decline of 57,000 over the past year.

This week also saw good productivity data, which some economists see as the first sign that companies are starting to use AI effectively.

So, companies may soon be able to do more with fewer people.

But they might also just be doing more.

A new paper in Harvard Business Review found that “AI doesn’t reduce work; it just makes work more intense.”

In an eight-month study of work practices at a tech company, the authors found that AI accelerated work pace, expanded task scope, and extended work hours into more parts of the day.

“Many people send prompts to AI while eating lunch, in meetings, or waiting for files to load. Some describe sending a ‘last quick prompt’ before leaving their desk, so AI can keep working while they’re away.”

For employers eager to extract more value from employees, this sounds appealing. And even better: “Employees are increasingly absorbing tasks that previously would have required additional staff or headcount.”

But researchers warn employers:

In the short term, seemingly higher productivity may mask the silent expansion of workload and growing cognitive pressure, as employees handle multiple AI-driven workflows simultaneously. Because extra effort is voluntary and often described as “fun experiments,” leaders can easily overlook how much additional burden employees are actually taking on. Over time, overwork damages judgment, increases errors, and makes it harder for organizations to distinguish real productivity gains from unsustainable work intensity.

If that’s the case, companies may soon find they need more people, not fewer.

At least, IBM’s HR chief expects so. Nick LaMoreaux told Bloomberg that cutting early-career hiring might save money in the short term but could lead to a shortage of mid-level managers later.

So IBM plans to double its entry-level hiring. “Yes,” LaMoreaux said, “specifically for those jobs everyone says AI can handle.”

The investment bank I worked at was constantly hiring between rounds of layoffs—replacing staff as they tried to figure out who was doing what.

The entire U.S. economy might soon do the same.

Let’s look at the chart.

This morning’s employment report was “brutal” for the tech sector. Losing 57,000 jobs over the past year, “almost as bad as the worst period of 2024 tech downturn, and clearly more severe than the recessions of 2008 or 2020.”

The tech industry is just the tip of the iceberg. Looking at the entire U.S. economy, according to global re-employment and executive coaching firm Challenger, Gray & Christmas, employers announced layoffs of 48,307 workers in February. This is a 55% decrease from the 108,435 announced in January, and a 72% drop from 172,017 in the same month last year.

In January and February alone, a total of 156,742 layoffs were announced—making it the lowest start to the year since 2022 (when only 34,309 layoffs occurred in the first two months). But even so, this number ranks fifth highest among the same period since 2009.

In other words: the layoff wave has indeed slowed compared to early this year and last year, but in a historical context, it’s still high. Workers’ days of easy employment are not coming back soon.

Too many leaders?

An academic paper found that generative AI is creating a “credential-based technological shift” in employment, with particularly severe impacts on entry-level workers. This isn’t just happening in tech: the study analyzed resume data from 285,000 employers.

Hiring slowdown:

The same study explains that the decline in entry-level jobs is “entirely driven by a decrease in hiring.”

AI effects:

Websites long used for buying advice, like Wired and Tom’s Guide, have seen traffic plummet. Now we go straight to chatbots—

And the sources of information these bots draw from are the very sites they’re pushing out of the market.

Or is it AI?

AI professor Alex Imas points out that this week’s productivity data “shows signs” that companies are already benefiting from AI.

Just talk?

Goldman Sachs (via Callum Williams) reports that although 70% of companies are talking about AI, only 10% can explain how it helps their business, and just 1% can quantify its impact on profits.

Work is always changing:

Tech journalist Roland Munsop mapped out the most common jobs in the 1980s and found that “secretary” was the most common job in 19 U.S. states.

What AI can and cannot do:

Peter Walker reorganized data from Anthropic showing which parts of each profession AI theoretically can perform (blue) and how much they are actually performing now (red).

Great question!

In a reply on X platform, Boris Cherny, responsible for Claude Code, explained that all the code Claude is writing is creating new jobs that only humans can do.

That’s a great job if you can get it:

Annual salary: $405,000–$485,000.

These are some of Anthropic’s open positions and their salaries. The code is writing code, but someone still has to tell it what to write—and that’s a high-paying job.

Claude is winning:

A stunning chart from Ramp shows OpenAI’s shrinking share (blue) versus Claude’s growing share (orange) in the commercial market.

Time mismatch:

A Gartner study predicts that “AI will not bring an ‘employment end-of-days’—but it will cause employment chaos.” They expect that starting in 2028, jobs created by AI will outnumber those eliminated.

Call me an “apocalyptic optimist,” but I think all this will happen faster than expected.

Wishing all hardworking readers a happy weekend.