Meta CTO Admits AI Reorganization Was 'Atrocious,' Pledges Management Overhaul
Key Takeaways
- ▸Meta's rapid March 2026 reorganization of its Applied AI division (6,500 engineers) was poorly communicated and executed, with employees describing work as menial and comparing it to 'a gulag'
- ▸CTO Andrew Bosworth publicly acknowledged management's failure to explain strategy, support career growth, and maintain organizational stability amid rapid hiring and strategy shifts
- ▸Meta is implementing governance changes: capping direct reports at 20, limiting manager changes, and introducing AI coaching tools to rebuild employee trust
Summary
Meta's chief technology officer Andrew Bosworth acknowledged that the company's reorganization of its Applied AI division in March was executed poorly, undermining employee trust and contributing to widespread internal dissatisfaction. The reorganization created a 6,500-person team focused on improving Meta's generative AI models, but employees complained about menial work assignments and lack of career clarity, prompting Bosworth to take accountability for management's failure to communicate the strategic vision.
In an internal memo addressing the discontent, Bosworth attributed the missteps to executives losing sight of employee perspectives while rushing to compete in the AI market. He acknowledged: "We obviously did an atrocious job explaining the vision, giving people a clear picture of how we would support them and their careers in the shift." The public admission reflects broader morale challenges at Meta stemming from recent mass layoffs, worker surveillance policies, and perceived management instability.
Meta is responding with structural changes designed to restore employee confidence. The company will cap managers at approximately 20 direct reports each, limit reassignments during restructurings, and provide AI coaching tools. Additionally, employees who were forcibly transferred to the Applied AI team can now transfer to other roles within Meta, effectively walking back the mandatory placement policy. VP Maher Saba framed the shift as a pivot from forced scaling to voluntary engagement, allowing the team to continue focusing on AI coding and agentic capabilities.
- Forced transfers to the AI team are now voluntary; employees can apply for other roles within Meta, effectively reversing the mandatory placement mandate
