Inside Meta's AI-Driven Layoff Revolution

A senior engineer at Meta is sounding the alarm from inside the company: even employees with strong performance reviews and no presence on layoff lists feel their jobs are on borrowed time. The engineer, identified by the pseudonym Eva, made a striking claim — Meta could lay off 90% of its workforce and Instagram, Facebook, and its other products would continue running normally.

This insider account, originally published by Chinese tech outlet Dongcha Beating, offers a rare and unsettling window into how artificial intelligence is fundamentally reshaping employment at one of the world's largest technology companies — not in some distant future, but right now.

Key Takeaways

• Meta has laid off approximately 25,000 employees since 2022, with 11,000 cut in November 2022 and another 10,000 in 2023
• A current senior engineer says 'nobody is safe,' regardless of performance ratings
• The engineer is actively embracing AI tools but still considers job loss 'a matter of when, not if'
• Meta's layoff rationale is increasingly disconnected from individual performance
• The nature of engineering work at Big Tech is being redefined by AI coding assistants
• Performance evaluation, promotion systems, and management structures are all under pressure

'The Layoffs Are Real, but the Reasons Are Fake'

Eva's most provocative claim cuts to the heart of a growing tension in Silicon Valley: the stated reasons for layoffs often have little to do with the actual motivations behind them. Meta framed its massive workforce reductions as part of a 'year of efficiency,' a phrase coined by CEO Mark Zuckerberg in early 2023. But according to Eva, the real calculus is simpler and more brutal.

The company's core products — Facebook, Instagram, WhatsApp, and Messenger — are mature platforms that require far fewer engineers to maintain than they needed to build. When you layer AI-powered development tools on top of that reality, the math becomes devastating for headcount.

This pattern isn't unique to Meta. Across Big Tech, companies are discovering that AI coding assistants like GitHub Copilot, Meta's own CodeCompose, and emerging tools powered by large language models can dramatically accelerate software development. The implication is clear: fewer engineers can accomplish what once required entire teams.

How AI Is Rewriting the Rules of Performance

What makes Eva's account particularly compelling is its focus on the systemic shifts happening inside Meta's engineering culture. Performance reviews, once the bedrock of career advancement in Big Tech, are becoming increasingly disconnected from job security.

Eva reports having solid performance ratings and actively integrating AI tools into daily workflows. Yet the sense of vulnerability persists. This paradox reveals something important: AI isn't just automating tasks — it's automating the justification for keeping people employed.

In traditional Big Tech performance systems, engineers are evaluated on metrics like:

• Impact: How much measurable value did your work create?
• Code quality: How robust and maintainable is your output?
• Collaboration: How effectively do you work across teams?
• Innovation: Are you pushing the technical boundaries?

But when an AI tool can generate boilerplate code, write tests, debug issues, and even architect simple systems, the bar for what constitutes 'impact' rises dramatically. Engineers who once earned strong reviews for reliable execution now find that reliability alone isn't enough — the machine can be reliable too, and it doesn't need health insurance.

The Promotion Ladder Is Collapsing

Eva's testimony also sheds light on how promotion pathways at Meta are being disrupted. In the traditional Big Tech career ladder, engineers progress from junior roles (E3/E4 at Meta) through senior (E5) and into staff-level positions (E6 and above). Each rung requires demonstrating increasing scope and leadership.

But AI is compressing this hierarchy. When a junior engineer equipped with advanced AI tools can produce output comparable to a mid-level engineer from 3 years ago, the entire calibration system breaks down. Why promote someone to E6 when an E4 with Claude or GPT-4 can handle the same workload?

This compression creates a particularly dangerous dynamic for mid-career engineers — the very people who form the backbone of most engineering organizations. They're too expensive to justify purely on output, but they haven't yet reached the strategic altitude where their judgment and architectural vision become irreplaceable.

The numbers tell a stark story. Meta's engineering headcount peaked at roughly 77,000 total employees in late 2022. Today, after successive rounds of cuts, the company operates with approximately 67,000 — and by most accounts, its products have not suffered. If anything, Zuckerberg has publicly claimed the company is more productive than ever.

The Management Layer Faces Its Own Reckoning

It isn't just individual contributors feeling the heat. Eva's account suggests that middle management at Meta is facing an existential crisis of its own. In the 'year of efficiency,' Zuckerberg explicitly targeted management layers, flattening the organization and increasing the number of direct reports per manager.

AI accelerates this trend in several ways:

• Automated reporting: AI tools can generate status updates, track metrics, and flag blockers without human managers
• Decision support: LLM-powered systems can analyze project data and recommend resource allocation
• Communication synthesis: AI can summarize cross-team discussions and identify alignment issues
• Performance tracking: Automated systems can monitor code commits, review participation, and collaboration patterns

The traditional manager who served as an information router between teams and leadership finds their role increasingly redundant. What remains valuable is genuine people leadership — mentoring, culture-building, conflict resolution — but those skills are harder to measure and easier to deprioritize during cost-cutting cycles.

The Broader Big Tech Pattern

Meta's experience mirrors a trend playing out across the entire technology industry. Google has conducted multiple rounds of layoffs since early 2023, cutting over 12,000 jobs initially and continuing with smaller reductions throughout 2024. Amazon eliminated 27,000 positions across 2022 and 2023. Microsoft cut 10,000 jobs in January 2023.

What distinguishes the current wave from previous tech downturns is the AI factor. In the dot-com bust or the 2008 recession, layoffs were driven by revenue shortfalls. Today, many of these companies are posting record or near-record profits. Meta reported $40.1 billion in revenue for Q4 2024, up 21% year over year.

The layoffs aren't happening because the business is struggling — they're happening because AI makes it possible to maintain or increase output with fewer people. This is a fundamentally different dynamic, and it's one that traditional labor economics frameworks struggle to address.

Compared to previous technology shifts like cloud computing or mobile development, AI's impact on software engineering employment is moving faster and cutting deeper. Cloud computing created as many jobs as it displaced. AI, at least in its current trajectory, appears to be a net reducer of engineering headcount at mature technology companies.

What This Means for Engineers and the Industry

For software engineers currently working in Big Tech or aspiring to join, Eva's account carries several practical implications.

First, embracing AI tools is necessary but not sufficient. Eva is already using AI extensively and still feels vulnerable. The competitive advantage lies not in using AI, but in doing things AI cannot — understanding ambiguous requirements, navigating organizational politics, making judgment calls with incomplete information, and building relationships with stakeholders.

Second, the traditional career playbook of joining a big company, climbing the ladder, and retiring comfortably is increasingly unreliable. Engineers need to think about building portable skills and personal brands that transcend any single employer.

Third, the startup ecosystem may benefit as displaced Big Tech talent seeks opportunities where individual contribution still matters at scale. Early-stage companies where one engineer's work can make or break a product offer a fundamentally different value proposition than maintaining features at a mature platform.

Looking Ahead: The Storm With No Clear End

Perhaps the most unsettling aspect of Eva's account is the uncertainty. As the engineer puts it, 'from Zuckerberg down to the newest junior engineer, nobody can say when this storm will end.'

The reality is that it may not end — at least not in the way previous cycles have. AI capabilities are improving rapidly, with each new model generation expanding the scope of tasks that can be automated. Meta's own Llama models, OpenAI's GPT series, and Anthropic's Claude are all advancing at a pace that consistently surprises even their creators.

What we're witnessing at Meta may be the early stages of a permanent restructuring of how software companies operate. The era of massive engineering teams building and maintaining consumer products could be giving way to a leaner model where small teams of highly skilled engineers leverage AI to accomplish what once required hundreds of colleagues.

For the tech industry's workforce, the message from inside Meta's walls is clear: adaptability isn't optional, performance reviews won't save you, and the transformation is already underway. The question isn't whether AI will reshape Big Tech employment — it's how dramatically, and how fast.

Eva continues to show up to work every day, write code, attend meetings, and collect a strong paycheck. But the confidence that once came with a senior engineering title at one of the world's most powerful technology companies? That, it seems, has already been laid off.