Meta's embrace of AI is making its employees miserable

Meta’s AI Push and Internal Culture

• Many see Meta’s AI turn as a continuation of a long pattern of top‑down, optics‑driven bets (e.g., metaverse, crypto) with weak execution and poor product–market fit.
• Commenters describe Meta’s culture as fear‑based and political: work is hoarded by favorites, others get “bullshit projects,” and layoffs are perceived as spreadsheet‑driven regardless of individual performance.
• Some argue employees are demoralized less by AI itself than by serial reorganizations, repeated layoffs, and the sense that only the AI narrative and “tokenmaxing” matter.

Token Tracking, Surveillance, and Goodhart’s Law

• Internal dashboards tracking AI “token” usage are widely criticized as a dumb metric that encourages waste and gaming rather than productivity.
• Reports from other big companies echo this pattern: usage spikes after dashboards launch, but measurable output doesn’t improve; now people hoard tokens or time work around monthly budgets.
• Several warn of Goodhart’s Law: once a metric becomes a target, it ceases to be useful, and here mainly drives higher cloud bills and anxiety.

Technology, Power, and Capitalism

• A major thread questions the belief that technology inherently makes life better; instead, tech is framed as amplifying existing power structures and inequality.
• Some emphasize that the problem is capitalism/wealth concentration, not “technology” itself: automation could reduce work hours broadly, but instead enriches a small elite.
• Others argue certain technologies decentralize power (encryption, contraceptives, printing press), while LLMs and platforms like Meta are strongly centralizing.

AI in Everyday Knowledge Work (“AI Slop”)

• Many describe a new workplace frustration: colleagues use LLMs to generate long, low‑quality emails, docs, tickets, and PRs that take far more time to review than they took to produce.
• “AI slop” is seen as perfect for people optimizing for appearance of productivity; strong review cultures can detect this, but weak ones will be overwhelmed.
• Emerging informal norms: using AI for proofreading and drafting is fine; copy‑pasting unedited outputs into human communication is viewed as disrespectful and wasteful.

Open Source, Decentralization, and Regulation

• There is extensive debate on whether open source/“free software” has mitigated or accelerated corporate power concentration.
• Some argue open models and volunteer efforts could eventually counter centralized LLMs; others point to resource barriers (compute, data, curation) and corporate capture of open tech.
• Several call for “common‑sense guardrails” and regulation (disclosure, usage rules), while others claim strong regulation or trying to “prevent” AI progress is unrealistic.

Worker Response and Unions

• A subset pushes for unions or new collective structures, noting big‑tech workers are now experiencing the same precarity and metrics‑driven management long common elsewhere.
• Others are skeptical, but there is clear concern that AI is being used as a pretext to squeeze more output from fewer workers while eroding autonomy and morale.