Is Meta destroying its engineering organization?

That framing mostly held up. People with Meta experience said the story matched what they had seen, especially the split between stronger acquired orgs and weaker homegrown ones, and the long-running habit of scaling headcount around executive enthusiasms without a clear operating plan. Several commenters also said the forced labeling claims sounded believable because post-training quality is one of the few places where a giant company can still throw bodies at the problem. If leadership thinks coding models are the most direct path to revenue, moving expensive but already-vetted engineers into review and preference-labeling work is irrational for morale but not irrational for a panicked executive team. The bigger conclusion was that Meta looks less like an isolated mess and more like an extreme version of something spreading across tech. Mature ad platforms still mint cash, but they no longer offer enough product novelty to absorb giant engineering orgs. That leaves leaders chasing AI, VR, and other “next platform” bets while quietly acting as if core software is a solved problem and large parts of engineering can be repurposed, measured by tokens, or pushed out. Many readers saw the labeling reassignments as both a catch-up tactic and a soft layoff mechanism. Others went further and argued this is what happens when a company with monopoly-like economics, weak accountability, and a founder in permanent search of the next growth story stops knowing what its engineers are for. The mood was unsparing. There was little sympathy for Meta as a company, but a lot of concern that the same management pattern could spread elsewhere. The sharpest practical point was simple: AI may change software work, but the companies learning useful habits will be smaller teams with clear accountability. The ones turning it into leaderboards, forced compliance, and giant internal labor shifts are advertising their confusion, not their advantage.