Meta's work culture

credits to : the pragmatic engineer

Meta has drastically transformed its historically celebrated, engineer-centric culture into a rigid cost center due to an aggressive, top-down push for AI dominance. Driven by CEO Mark Zuckerberg and Scale AI CEO Alexandr Wang—following Meta's $14.8 billion stake in Scale AI—the company implemented highly controversial strategies to rapidly train its upcoming large language models.
First, Meta mandated invasive keystroke and mouse-click tracking for employees to generate AI training data. Second, leadership forcefully reassigned 30% to 50% of core engineering teams—roughly 4,500 to 5,000 software engineers—to the Agent Data Optimisation (ADO) unit for repetitive data labeling and reinforcement learning from human feedback (RLHF).
This internal disruption coincided with an announced 10% staff layoff and a hyper-aggressive performance review system that began evaluating AI token usage. Consequently, engineers prioritized "tokenmaxxing"—inflating their personal metrics using AI tools—over code quality and rigorous human review.
The severe understaffing of core infrastructure and security teams culminated in a historic outage on May 30. Flawed, AI-generated and AI-reviewed code introduced a zero-authentication password reset vulnerability, allowing hackers to easily hijack high-profile Instagram accounts. Meta's Chief Information Security Officer, Guy Rosen, resigned immediately following the incident.
Internal morale has collapsed, with employees describing the environment as a "gulag" and Chief Product Officer Chris Cox acknowledging the internal "insanity". While leadership is now attempting to scale back tracking and improve communication, the aggressive shift has prompted a massive wave of long-tenured talent to actively seek external jobs.

Видео Meta's work culture канала deepak lodhi