When the press releases share a script, that's the news

TL;DR

Snap, Meta, Oracle, Atlassian, Salesforce, Block and ASML have all run sizeable workforce reductions in 2026 with AI cited as the named reason.
The CEO memos and internal notes are converging on near-identical phrasing — Business Insider has called it AI-layoff "Mad Libs."
Polymarket is pricing the "Tech Layoffs Up or Down 2026?" market at roughly 85% YES on the up side as of 24 April.
The script is doing two jobs at once: it justifies cuts that would otherwise be hard to defend, and it sets the template for the next round of cuts at the next company.
The thing to do about it isn't to panic. It's to read your own employer's communications carefully and match the words against the actions.

What happened

In the first three weeks of April 2026, seven major employers announced workforce reductions of meaningful size. The numbers, in order: Snap (~1,000), Atlassian (1,600), Salesforce (~1,000), Block (4,000), ASML (1,700), Oracle (somewhere in the 20,000–30,000 range across multiple business units), and Meta (a roughly 10% reduction of about 8,000 people, scheduled for 20 May but reported on 22 April).

Most of these cuts named AI as the proximate cause. Read the language closely, though, and a stranger pattern emerges. The memos rhyme.

Business Insider's Pranav Dixit wrote it up on 16 April as "AI-layoff Mad Libs" — the corporate version of the party game where the words change but the structure doesn't. The skeleton sentence reads, with light variation:

"Over the last [single-digit] months, AI's impact has [adverb of speed] evolved from [potential / promise / hype] to [actionable / operational / real]…"

Fill in the blanks with your favourite Big Tech CEO. The phrasing has now appeared, almost verbatim, in communications from at least four of the seven companies above. It is a script.

What it actually means

A script means the language is doing rhetorical work that the underlying business logic isn't quite carrying on its own.

There are three things going on at once and only the first one is the AI story.

One. AI is genuinely shifting how some of these jobs get done. Software engineering is the clearest case. Customer support is the second clearest. Marketing copy production is the third. Inside those functions, the productivity gains are real, the headcount math is real, and a smaller team can plausibly produce more output. Some of these cuts are exactly what they say they are.

Two. AI is the cleanest excuse available for cuts that have other causes. Several of the announcements above sit on top of multi-quarter operating-margin pressure, post-ZIRP correction in tech-sector hiring, and slower-than-promised growth in the actual AI products being sold. "We are reorganising around AI" is a more flattering phrase than "we over-hired and the new product line hasn't lived up to the deck." The first one positions you as forward-leaning; the second one positions you as a bad operator. CEOs choose the first one.

Three. The script itself is now the product. Once a phrase has worked once, it works easier the second time. The Meta memo cribs from the Salesforce memo cribs from the Oracle memo. Internal communications teams read the press coverage. PR firms are now circulating decks. The vocabulary is hardening into a corporate dialect, and that hardening matters because it removes the one piece of friction that used to make companies pause before this kind of announcement: the discomfort of having to find your own words for it.

The hype deconstruction

The cleanest test of whether AI is the cause or the cover is to read what each company's most recent earnings call said about revenue from AI products, not spend on AI products. The two are different. Companies cutting on the strength of AI-revenue growth are making a coherent business call. Companies cutting on the strength of AI-spend rationalisation are doing something else and using the same vocabulary.

Run that test on the seven companies above and the picture splits roughly in half. About three of them have a defensible internal AI revenue story to point to. About four don't, and their cuts read, on the numbers, as ordinary cost reductions in a year of compressed margins. The script doesn't distinguish between the two. That is the rhetorical work the script is doing.

The Polymarket number — 85% YES on tech layoffs increasing in 2026 — is interesting not because it's a forecast, but because it's a sentiment reading. The market is saying: we expect the script to keep working.

Stakeholder landscape

Workers at scripted companies. Reading the script means reading what your employer is actually doing, not what it is saying. The named reason and the actual cause are not always the same thing, and the gap between them is the most useful piece of information you have.
Workers at companies that haven't run cuts yet. The Mad Libs structure is a leading indicator. If your CEO's all-hands deck has started using phrases that sound like the ones above, that's a signal worth taking seriously — about timing, not certainty.
Senior leaders. The script offers short-term cover and creates a long-term legibility problem. Once your workforce knows the phrasing, every future communication is read against it. Companies that find their own language for restructuring will retain more trust than companies that copy the template.
Investors and analysts. The Mad Libs pattern is a calibration tool. AI-stated reasons should be checked against AI-stated revenue contribution, not AI-stated investment. The two are diverging.
Workers in functions least exposed to LLM substitution today (clinical, trades, regulated professional services, hands-on operations, complex sales). The Mad Libs cycle is concentrated in software, support and content. The cycle has not yet credibly arrived in these fields. That is data, not a guarantee.

Cross-layer implications

Workforce. Job-loss anxiety is now decoupled from individual-performance anxiety in a way it wasn't five years ago. Workers in good standing are losing roles. That changes how people behave at work — quieter, more guarded, less candid feedback flowing upward. That has knock-on effects on product quality and decision speed within twelve months.
Hiring. Junior roles in the affected functions are being hit hardest. The pipeline implication is structural — if you don't hire a 2026 cohort, you don't have a 2030 cohort.
Capital allocation. Companies citing AI as the cause of cuts are also raising AI-infrastructure spend. The two motions together compress the operating cost base in one direction and inflate it in the other. That math is finite.
Culture. The script is becoming the new explanation for everything. When everything has the same explanation, no explanation is true.

What this means for you

If you work at one of the named companies — read your own employer's words against its earnings disclosures. Look at the gap between AI-spend language and AI-revenue language. The bigger the gap, the more the script is doing the work the business case isn't.

If you work at a company that hasn't yet run cuts — three things to watch for in the next 60 days. First, all-hands language. If "evolving from potential to actionable" or close paraphrases start appearing, that's a 90-day timeline indicator. Second, hiring freezes that are described as "rebalancing." Third, mid-level redundancies grouped by function rather than by team — the AI-reason cuts tend to flatten functional middle layers first.

If you're early-career — the cuts are concentrated where you are. The two responses that matter are: become unusually good at the thing AI is least good at on your team (judgment under ambiguity, stakeholder navigation, written reasoning across long contexts, in-person trust-building), and build a portable network outside the company before the company tells you to. Both are slow. Start now.

If you lead a team — find your own language. The script saves you communication time today and costs you trust capital tomorrow. Workers can tell when they're being addressed in a dialect rather than a sentence. Plain words about plain trade-offs land better than scripted language about transformation. Always have. Always will.

If you're a board member or investor — separate AI-revenue stories from AI-spend stories in your reading. Pattern-match the words against the numbers. The companies that come out of this cycle in best shape are the ones whose AI narrative is doing one job at a time.

Uncertainty ledger

The cuts that happen between now and the year-end will sharpen the picture. If the script keeps working, the pattern compounds. If a company cuts under different language and the market rewards it, the script breaks.
Productivity data on actual AI-augmented teams is still thin. The headcount-substitution math at the cut companies will be readable in twelve to eighteen months when output quality and product velocity become measurable.
Whether AI displacement is a permanent floor lowering or a short period of overshoot followed by re-hiring is genuinely unresolved. Both have happened in past technology shifts. We will not know for two years.
The legal and disclosure exposure of the script itself is not yet tested. If a securities case turns on whether AI was the real reason for cuts that were actually about margin pressure, the language could become a liability rather than a defence.

The bottom line

The cuts are real and some of them are about AI. The script is real and most of it isn't. When seven major employers reach for the same handful of phrases inside one month, the phrases are no longer reporting — they are performing. The most useful thing a worker, a leader, an investor, or a journalist can do this quarter is to stop reading the memos as memos and start reading them as a dialect. Match the words against the numbers. Listen for the cliché. Notice when your own employer's language starts to rhyme with someone else's. The script tells you something. It just doesn't always tell you what it says.

Sources

Business Insider, AI-layoff Mad Libs (Pranav Dixit), 16 April 2026 — Tier 1
Reuters, Snap, Atlassian, Salesforce, Block, ASML layoff filings and statements, April 2026 — Tier 1
WSJ, Meta to cut roughly 10% of workforce on 20 May, 22 April 2026 — Tier 1
IBT UK, Oracle "stretch" reorganisation reporting, 1–2 April 2026 — Tier 2
Polymarket, Tech Layoffs Up or Down in 2026?, market data captured 24 April 2026 — Tier 2
Various CEO memos, all-hands transcripts, internal communications quoted in the above — Tier 2 (paraphrased / leaked)