Atlassian just laid off 1,600 people, 10% of the company. The stated reason: reallocating resources toward AI.
A few weeks ago, Block did almost the exact same thing. Jack Dorsey said AI could replace a lot of roles. Now Atlassian follows with nearly identical language. I went through their public statements, and the pattern is consistent: management talks about being "AI-first," but which teams were cut, and what the actual AI product roadmap looks like, all vague. Nobody says whether the AI can actually do what the people who just left were doing.
This is becoming a new kind of layoff script: tell investors you're going all-in on AI, tell employees thanks for your service. It's a framing that's almost impossible to argue with, because the AI future is always just around the corner, specific enough to justify the cuts, vague enough to avoid accountability.
Meanwhile, the consumer side had an unusually dense week of AI feature launches across social apps.
Bumble rolled out an AI dating assistant called Bee, no more swiping left or right, it now matches based on compatibility signals. Facebook Marketplace is letting sellers use Meta AI to auto-reply to buyer messages. Tinder is running in-person events combined with virtual speed dating, using AI to win back younger users who've stopped using the app.
Three major platforms, same week, same bet: AI handles the talking, humans just show up. That's not a coincidence.
On the capital side, the pace is just as aggressive.
Wonderful just raised a $150M Series B at a $2B valuation — four months after closing a $100M Series A. The valuation doubled in four months. Even by today's frothy standards, that's fast. Separately, Rox AI, a sales automation startup founded by the former head of growth at New Relic, raised at a $1.2B valuation. Their pitch: replace legacy CRM with AI-native tools.
The throughline is pretty simple: money is moving toward anything that cuts humans out of the loop.
One more thing, and this one's different.
Journalist Julia Angwin has filed a class action lawsuit against Grammarly, alleging the company used users' writing to train AI models without consent. The claims rest on privacy and likeness rights. Angwin is a serious investigative tech journalist — the fact that she's the one bringing this case suggests that the practice of silently feeding user content into model training has reached a point where even people inside the industry have had enough.
AI isn't just replacing blue-collar or white-collar work — it's replacing the communication between people. Your buyer messages get handled by AI. Your matches get decided by AI. You got laid off because your company needs to fund AI. And the tool you use to write your emails might be training the next model on everything you've written.
The question that's worth sitting with: in 2026, are you an AI user, or are you AI's raw material?