It takes a particular kind of dysfunction for an employee to hijack a company-wide livestream and demand the hosts relay a message to a senior executive: that he is, in so many words, a piece of garbage. That happened inside Meta this week. One of the presenters reportedly covered their face with their hands. Then everyone was asked to mute, and the technical talk carried on. The outburst, first reported by WIRED, was not really about one executive. It was the loudest signal yet of how badly things have soured inside Applied AI, the 6,500-person unit Mark Zuckerberg built barely three months ago to power his most expensive bet in artificial intelligence.That bet has a name: Alexandr Wang. Zuckerberg paid $14.3 billion last summer for half of Wang’s data-labeling startup, Scale AI, and installed the young founder as Meta’s chief AI officer—the company’s highest-paid hire. Applied AI was supposed to be the machinery behind him, the team that turns Wang’s models from promising to dominant. The engineers assigned to that machinery are using a different vocabulary. Soul-crushing. The gulag. A total mess.
The job nobody applied for
The complaints start with the work itself. Where these engineers once shipped features for billions of Facebook and Instagram users, they now spend their weeks generating coding problems and puzzles to train Meta’s models—sometimes a quota of two a week. One described the task as mechanical and creatively dead. “You have zero purpose in life all of a sudden,” another told WIRED. What deepens the resentment is how they landed here in the first place. The Wall Street Journal reported the unit’s creation back in March, noting its strikingly flat shape: as many as 50 engineers to a single manager. By April, an internal memo from unit chief Maher Saba—seen by Reuters—had closed the exits. Transfers, he wrote, were not optional. The people moved against their will began calling themselves draftees.
Meta layoffs, worker surveillance and a morale freefall
None of this is happening in a vacuum. Meta cut about 8,000 jobs in May—a tenth of its workforce—after quietly shifting another 7,000 people onto AI teams two days prior. Employees pushed back where they could. More than 1,600 signed a petition against a program that monitors their clicks and keystrokes to harvest AI training data; office flyers nicknamed it the “Employee Data Extraction Factory.” The salad emoji became internal shorthand for a salute to the laid-off.
Can Zuckerberg’s superintelligence gamble survive its own staff?
Even Meta’s leadership has stopped pretending otherwise. CTO Andrew Bosworth conceded in a memo, also obtained by WIRED, that the rollout was “atrocious.” Zuckerberg admitted to “mistakes,” promised no further mass layoffs this year, and moved to shrink those bloated manager ratios. The apology tour tells you how much is riding on this.And plenty is. Meta’s stock has slid 18 percent over the past year, the weakest of the tech giants. Wang’s debut model, Muse Spark, arrived in April to a collective shrug. Zuckerberg needed Applied AI to prove Meta still belongs beside OpenAI, Anthropic and Google. What he has instead is a flagship team that describes itself as conscripted labor—and says so on a livestream, for thousands of colleagues to hear.