The Room Is Quieter Now

It’s 11 PM. You’re still at your desk.

The fifth round of layoffs this quarter went out last Friday, and your name was not on the list. Neither was the name of the person you sit next to. At least not that round. Two from your team, though. One of them had a kid starting school next month. The other had just closed on a house.

You’ve been doing the math in your head since 5 PM. Rent. Runway. The credit card that you were going to pay off this quarter. What you’d say to your partner if the next round took you. What you’d say to your kids. Whether the new hire joining next week—you know… the tool-fluent, over-keen, half-your-age, twice-as-fast-at-spinning-up-prototypes—is your replacement on a twelve-month delay.

You close the laptop. But you don’t feel anything specific. Just tired. In a way that has no name. Like you’re “done” but don’t know what the next “to-be-done” is. You tell yourself, like you’ve been telling yourself for months, that you’re lucky. That you still have a job. That other people have it worse. That you should be grateful. (Sounds like Anne Hathaway from the opening of the The Devil Wears Prada 2. And if that movie has points—and it has several—to make about the current cultural zeitgeist, well… we all fucked ourselves over pretty good now, didn’t we?)

You try to believe this. You get into bed. But you don’t sleep.

If this is where you are right now, if any of this is yours, stay with us. This piece is for you. We are not going to tell you how to fix it. We are not going to tell you what to do. We are going to tell you, as clearly and as carefully as we know how, that what you are feeling is real. That you are not imagining it. That you are not alone in it. And that the weight you have been carrying quietly is, in fact, heavy.

The Next Morning

The next morning you go into a meeting.

You’ve been in this kind of meeting for fifteen years. Maybe twenty. You know the shape of the room, the rhythm of the agenda, the way the real decisions get made in the five minutes after the Zoom recording stops. The people around the table have, for most of your career, deferred to your read on things. But because you had earned it. Through toil and turmoil. Earned it.

Today the founder, or the VP, or the new hire they promoted in December, brings up a demo. Something they built last night in a tool you had heard of but not used yet. They run the prototype. It looks fine. It is, if you are honest, for all intents and purposes, just that: fine. It’s not better than what your team spent six weeks iterating on. But it is… there. On the screen. Right now. And the six weeks of iteration is on a Figma board somewhere that nobody in this room has opened today.

“Why don’t we just go with this,” they say. Or: “This is actually pretty close to what we need.” Or, and this is the one you’ve started hearing so often that you could set your watch to it: “The nice thing about AI is that we can just prototype it and save all the back-and-forth.”

You say something. A note, a concern, a question about the decision logic underneath what they just showed. And your voice sounds, to your own ear, slightly pre-apologetic.

You didn’t use to sound like that.

The room nods politely. The conversation moves on. The decision has already been made. You are still technically the senior person in the room. You are also, as you realise somewhere between the third and fourth agenda item, no longer leading.

You are not being fired. That is the cruel part. You are being quietly “un-heard”. One meeting at a time. In a pattern that is so slow and so plausible that you could never point to a single moment and say, “this… is where it happened”. When you try to describe it to your partner over dinner, you can’t find the words. You end up saying “it’s fine, just a weird day”, and changing the subject, because the alternative is admitting that something you have spent two decades building is being unbuilt by people who think a prototype is the same thing as a sound judgement. Intention. Taste.

You go back to your desk. You check your work three times before sending it.

You didn’t use to do that either.

You Are Not Alone

If you recognise either of these rooms, the first one or the second: the survivor at their laptop or the senior still in the chair but no longer in the decision, you are not alone in them. You feel alone because the rooms are designed—sometime accidentally, mostly intentionally—to isolate. Everyone in the first room is doing their own private math and assuming that the person next to them is fine. Everyone in the second room is force-swallowing their own observation, and their pride, and assuming the problem is them.

That aloneness is a part of what is being done to you.

The research agrees. Classen and Dunn, in Health Economics, found one additional suicide death for every 4,200 men laid off in a mass-layoff event, and one for every 7,100 women. A 2023 systematic review pins long-term unemployment to rates of diagnosed depression as high as 50%. The numbers carry their own weight. We are not going to embellish them.

The part that almost nobody in the AI-layoff discourse names: the research on survivors is nearly as severe as the research on the people who were let go. The people still in the building, absorbing the departed colleagues’ work, watching the next list take shape. Belle and Bullock document psychological distress in this group at levels comparable to those who were actually let go. A Swedish cohort study went further: perceived job insecurity, independent of actual job loss, was associated with increased suicidal behaviour.

The thing you are feeling at your desk at 11 PM is not background anxiety. It is, in the research literature, a measurable mortality-adjacent risk factor. You are not imagining it. Your body knows. Your body is correct.

Seven Hundred People

Now we can talk about the decisions that produced these rooms.

In early 2024, Klarna’s CEO Sebastian Siemiatkowski announced, publicly and with some amount of swagger, that the company’s AI chatbot (built in partnership with OpenAI) was doing the work of 700 customer service agents. Hiring was frozen. The workforce shrank by roughly 22% through attrition. In January 2025, Siemiatkowski posted on X: “AI can already do all our jobs.”

By May 2025, Siemiatkowski was telling Bloomberg that the AI-first strategy had “led to lower quality.” The company began rehiring. By October 2025, after the company’s successful US IPO, with shares surging 30% and valuation hitting $19.65 billion, Siemiatkowski formally admitted the aggressive AI cuts had “led to diminished service quality.” The final version of the admission, given to the press: “We focused too much on efficiency and cost. The result was lower quality, and that’s not sustainable.”

Seven hundred people lost their jobs. So one CEO could test whether he was right.

He wasn’t.

That’s not rhetoric. It is accurately descriptive of the situation though. A hypothesis was tested against the lives of 700 specific people. The hypothesis was proven wrong, and the 700 lives are not being refunded. They can’t be. The rehiring targets students, rural workers, and loyal Klarna customers. It does not target, and cannot target, the specific people whose specific livelihoods were used to discover that the AI was not, in fact, ready. Those lives—the mortgages, the kids’ schools, the career arcs, the eighteen months of unemployment-duration clock already ticking—do not get a do-over.

As we’ve argued before, the hardest decisions in a company are always the ones about what you actually know versus what you’re willing to claim to know. Klarna skipped the hard decision about whether their AI was ready. They let the market, via eighteen months of declining customer satisfaction, decide for them. The 700 paid the price for that skipped decision, in a currency the company wasn’t paying with.

Eighty Thousand Kitchens

And while Klarna is the clearest example, it is not the largest.

In Q1 of 2026 alone, nearly 80,000 tech workers were laid off globally. The TrueUp tracker logged 171 separate events (59,121 workers) at an average of 704 jobs a day. If that pace holds through the year, total 2026 cuts could reach 265,000. We are running ahead of 2025, and 2025 was already a record.

The stated reasons, compiled from the official announcements, are a masterpiece of corporate euphemism:

Amazon eliminated 16,000 roles in 2026 on top of 14,000 in late 2025. Roughly 30,000 jobs since October. The company reported $716.9 billion in 2025 revenue, a record, and $35 billion in profit in the first half of 2025 alone. The public framing: “flattening management layers.” Thirty thousand people were apparently management layers.

Meta cut 3,600 employees, approximately 5% of its workforce, in early 2025. Mark Zuckerberg cited, as the key reason, “raising performance standards and cutting low performers.”

That sentence deserves its own line.

Mark Zuckerberg cited, as the key reason for the layoffs, raising performance standards and cutting low performers.

You can do the translation yourself.

Microsoft cut approximately 6,000–7,000 roles in May 2025. Its largest reduction since 2023, made while announcing $80 billion in AI infrastructure investment in the same quarter. Google continued trimming hundreds of roles across Platforms and Devices, Pixel, Android, Chrome, and ad sales, framed as a move to “prioritise AI projects.” Block, under Jack Dorsey, reduced its workforce by nearly 40%, citing AI-driven efficiency and flatter teams.

The same IBTimes report cites RationalFX’s analysis attributing roughly one in five of the 2026 layoffs directly to AI adoption. The other four-fifths are attributed to overstaffing, restructuring, cost optimisation, flatter management, efficiency. We will leave the reader to wonder why so many unrelated efficiency imperatives are arriving in the same quarter, with the same framing, pointed at the same seniority brackets, during the same AI cycle. That particular interrogation is the next piece in this series.

For now, the relevant fact is narrower: regardless of the stated reason, the people in those 80,000 seats in Q1 have houses, kids, mortgages, ageing parents, medical bills, and lives that do not pause because a quarterly earnings call needed a coherent narrative. The stated reason is on the earnings call. The cost is in the 80,000 kitchens where, right now, somebody is doing the math.

The Slow Invisibling

Here’s the part the discourse keeps missing, and almost nobody is writing it cleanly.

The violence of this moment is not only the layoffs. The violence is also the slow invisibling of the people who were not laid off. The senior leader in the second room. The one whose experience used to count for something. The one who is now, quietly, being demoted in every meeting without being demoted on paper.

As we’ve argued before, taste is a scar, not a style. It is what you have after a career of being wrong, being right, being ignored, being corrected, being proven out, being rebuilt. It’s not downloadable. It can’t be prompted. It lives inside a specific person with a specific sum of all their life experiences.

In a well-functioning organisation, that scar is the most valuable thing in the room. In an organisation caught in The Blanding though, it is a liability. Because the person carrying the scar is slower than the prototype. More cautious than the demo. More complicated to manage than the over-keen, eager-to-please, junior with the tool. The senior costs more. Asks harder questions. Insists on the uncomfortable two-week investigation before the one-hour decision. From a velocity perspective, the senior is, in the bluntest terms, a drag.

So the senior is not fired. The senior is simply, slowly routed around.

But the memos call it “flattening”. Or “modernising”. Or, our personal favourite: “unlocking velocity”. What’s actually being unlocked is the company’s ability to skip the part where someone with two decades of scars says, “hold on, I’ve seen this shit before, and it doesn’t work the way you think it does”.

The cost of this pattern is not visible in any layoff tracker. It is visible in the senior’s body, at the desk, at 11 PM. In the senior’s partner noticing they’ve stopped talking about work. In the senior checking their work three times before sending, because they have lost the ambient certainty that their judgement still has a chair in the room. Six months later, the senior leaves on their own terms. Or, worse, stays and adapts to a smaller version of themselves, because the alternative is the first room… the one with the math.

A recent industry analysis from Muzli named an adjacent piece of this with unusual clarity: the apprenticeship model of senior craft is breaking, because the entry-level tasks that used to build judgement have been compressed by AI generation. The junior who never had to struggle with the first draft will never develop the muscle that makes the tenth draft precise. Which means, in ten years, the room that is now quietly overruling its seniors won’t have seniors at all. It will have a generation of grown-old-but-not-grown-up operators who were never given the friction that produces judgement, making decisions in rooms where the scars are gone.

And that’s not a dark prediction. It is the foreseeable, mathematical consequence of the choices being made right now, in meetings exactly like the one you were in this morning.

Back to the Desk

We want to return, one more time, to the first room. To the desk. At 11 PM.

If you are there tonight, or if you’re in the second room, the meeting-room version, the one where your authority is being eroded one agenda item at a time, we want you to hear this as clearly as we can say it.

You are not imagining it. The research agrees with you. The layoff trackers agree with you. The CEO who cut seven hundred people eventually agreed with you. The body that is tired at a depth you cannot explain is tired for reasons that are real, measurable, and documented in the peer-reviewed literature.

You are not failing. You are responding correctly to something that is actually happening.

We are not going to tell you what to do about it. Not in this piece. The constructive work lives later in this series: the part about what remains worth defending, about what still matters, about why your skills and your judgement and your hard-earned specificity are not obsolete and will not be made obsolete by the currently unraveling circus. It is coming. We will sit with you through it.

But before we can get there, honestly, we have to name a few more things. In the next piece, we are going to name the specific organizational cowardice that produces decisions like Klarna’s. The cowardice of shipping work you know is inferior, and pretending otherwise. After that, we are going to name what the decisions were actually about, which is not the thing the announcements said they were about. After that, we are going to name the economy of grift that is monetising your exhaustion, and why it is structurally compounding the violence of the layoffs themselves.

And then, finally, we are going to sit together. Not above you. Not beside the industry looking at you. With you. And talk about what is worth defending, and why the future belongs to the people who are still paying the cost of caring. That’s you. And we mean that the way the research means it.

For now: the math you’ve had to do was real math. The meetings you’ve had to walk out of were real meetings. The quiet is real. The feeling is real.

You are not imagining any of it.

You are seen. You are not alone. We are with you.

Stay with us.

Sources and references held while writing this

On the mental health and mortality research. Classen, T. and Dunn, R. (2012). The effect of job loss and unemployment duration on suicide risk in the United States. Health Economics 21(3): 338–350. PubMed and PMC versions linked inline. Moore et al. (2023). Mental health and unemployment: A systematic review and meta-analysis. ScienceDirect, linked inline. Belle and Bullock, The Psychological Consequences of Unemployment, SPSSI policy statement, linked inline. Blomqvist et al., Swedish cohort study on perceived job insecurity, PMC, linked inline. Unemployment, Behavioral Health, and Suicide. Health Affairs policy brief, 2022.

On the Klarna arc. Tech.co, May 2025. FinTech Weekly’s May 2025 coverage and SRN News’s November 2025 coverage of the post-IPO admission, both linked inline. LaSoft’s analysis of the rehiring announcement, linked inline. Siemiatkowski quotes drawn from Bloomberg interviews summarised in those sources.

On the 2026 layoff landscape. Tom’s Hardware, April 2026 — Tech industry lays off nearly 80,000 employees in the first quarter of 2026. IBTimes’s March 2026 reporting (TrueUp tracker data and RationalFX analysis), linked inline. The Workers Rights’ April 2026 Amazon coverage, linked inline. BizzBuzz on Microsoft, Google, and IBM cuts, and Outlook Business on the broader 2025 list, both linked inline. Zuckerberg quote per Meta’s 2025 layoff announcement.

On the apprenticeship-compression argument. Muzli’s April 2026 piece on what AI-generated UI means for senior craft, linked inline.

From the Methodborne archive. Taste is Judgement: Why AI Can’t Replace Earned Discernment and Why Most Early-Stage Startups Get Brand Strategy Wrong, both linked inline.