Meta's Applied AI Unit Faces Internal Turmoil

WIRED reports that Meta's newly formed Applied AI unit, created in March to support Meta Superintelligence Labs researchers, is experiencing acute internal dissatisfaction. The unit comprises about 6,500 engineers and product managers. A livestreamed employee meeting open to thousands was disrupted by an attendee urging others to 'tell him that he\'s a piece of shit,' with a presenter covering their face during the incident. Multiple current employees anonymously told WIRED they find their new assignments menial: one called the situation 'literally the gulag,' and another described the work as 'soul-crushing.' WIRED reports some staff were reassigned to generating puzzles to evaluate how reliably AI models from Meta and other companies can solve them. Meta declined to comment for the story.
What Happened
WIRED reports that Meta assembled a new Applied AI unit in March, combining roughly 6,500 engineers and product managers to support the work of AI researchers at Meta Superintelligence Labs. According to WIRED, the group has generated notable employee frustration. A livestreamed internal presentation open to thousands of employees was interrupted by an attendee who urged others to write to a specific Meta AI executive and 'tell him that he\'s a piece of shit,' with a presenter covering their face during the outburst.
Employee Accounts
WIRED quotes multiple current employees speaking on background who describe reassigned work as menial or demoralizing. One is quoted saying 'it\'s literally the gulag,' adding they have 'zero purpose in life all of a sudden' and just have 'tasks every week.' A second employee describes some of the tasks - including generating puzzles to test how reliably AI models from Meta and other companies can solve them - as easy compared to prior software development work. A third says 'most people find the work soul-crushing' and that 'almost all' employees seem unhappy. Meta declined to comment for the story.
Technical Context
Companies building centralized AI support organizations commonly use large cross-functional teams to scale model evaluation, data labeling, and benchmarking workflows. Evaluation tasks - including generating puzzles and edge-case prompts to probe model reliability - are a standard part of AI development pipelines, but these roles can feel like a significant step down for engineers who previously held product-facing software positions.
What to Watch
Track employee attrition rates, public recruiting listings for the Applied AI unit, and any follow-up statements from Meta. Watch whether Meta publishes technical documentation or tooling for the unit\'s evaluation tasks, which would signal investment in automating or scaling the work described by employees. Any public comments from Superintelligence Labs leadership on the Applied AI team\'s role would also clarify the unit\'s trajectory.
Scoring Rationale
The WIRED investigation documents dysfunction inside a 6,500-person team at one of the world's largest AI employers, with anonymous quotes from multiple current employees and a disrupted all-hands meeting. The story is notable to AI/DS/ML practitioners for workforce and talent-pipeline implications, but ranks as a workplace/organizational story rather than a technical or strategic development, placing it in the upper Notable range.
Practice with real Ad Tech data
90 SQL & Python problems · 15 industry datasets
250 free problems · No credit card
See all Ad Tech problems

