Meta spent 20 years building one of Silicon Valley's most respected engineering cultures. It started with the 'move fast and break things' era, codified in a 70-page internal book distributed to employees when Facebook hit one billion users in 2012, and matured into a 'move fast with stable infra' discipline in the early 2020s. Engineers were empowered, impact-focused, and treated as a profit center. That ended around April of this year.
In a matter of weeks, Meta leadership systematically dismantled that foundation. Engineers are being forcibly assigned to AI data-labeling tasks, software engineering has been reclassified from profit center to cost center, and the company suffered what insiders describe as its most embarrassing outage on record. The piece documents six distinct failure modes: AI mandate overreach, demoralized core engineering staff, internal organizational chaos, and a string of self-inflicted wounds that no external competitor caused.
The full article by Gergely Orosz is worth reading because the diagnosis goes beyond Meta. It asks whether 'AI psychosis', the compulsive subordination of engineering judgment to AI adoption metrics, is spreading across the industry. The answer has direct consequences for every engineering org currently being told to measure headcount reduction against AI tool usage.
[READ ORIGINAL →]