😺 Meta used staff as AI training data. Then cut them.
The Neuron · Eric Gerard Ruiz · 2026-05-21
Meta secretly monitored employee keystrokes and activity across Gmail, VSCode, and internal tools to train AI models, then laid off approximately 8,000 workers on the same day a leaked audio recording revealed Zuckerberg's true rationale for the surveillance program.
Extraction
Topics: employee-surveillanceai-training-datametaai-labor-displacementai-industry-news
Claims
- Meta installed keystroke and mouse-tracking software on employee computers and monitored activity across Gmail, GChat, Metamate, and VSCode to collect AI training data.
- Zuckerberg privately explained the surveillance as learning from 'really smart people,' a justification that diverged significantly from Meta's public framing of routine software usage monitoring.
- Meta laid off approximately 8,000 employees on the same day the leaked all-hands audio became public, having previously announced reassignment of 7,000 workers to AI-focused teams.
- OpenAI is preparing to confidentially file for an IPO targeting a September debut at an $852B valuation.
- The White House briefed major AI companies on a planned executive order allowing government agencies to review advanced AI models up to 90 days before public release.
Key quotes
You've heard of 'learn by doing.' Meta invented 'learn by watching employees do it, then replace them.'
The gap between the public story ('we're just training models to use software') and the private one ('we're learning from watching our best people') is the whole story.
The line between 'helping you work better' and 'training your replacement' just got a lot blurrier.