- Meta's CTO, Andrew Bosworth, admitted the company's AI reorganization was poorly communicated and that leadership did "an atrocious job explaining the vision."
https://www.inc.com/jessica-stillman/the-worst-its-ever-been-why-metas-massive-ai-reorg-backfired-spectacularly/91363370
Bosworth acknowledged employee morale has likely reached "the worst it's ever been."
The new Applied AI division left many engineers feeling directionless, unsupported, and disconnected from leadership.
One major problem was excessively large management spans, with some managers overseeing well over 20 direct reports.
Employees reported feeling like anonymous members of oversized, hastily assembled teams rather than valued contributors.
Meta now plans to cap managers at roughly 20 direct reports, reduce unnecessary manager changes during reorganizations, and introduce AI coaching tools.
Organizational psychologist Bob Sutton noted that research has long shown large teams suffer from coordination, collaboration, and communication problems.
Studies indicate that smaller teams are particularly important for creative, innovative, and highly interdependent work like AI development.
Research covering more than 50 million papers, patents, and software projects found teams with fewer than five members were most likely to produce disruptive innovations.
The article compares Meta's situation to Jeff Bezos' "Two Pizza Rule," which advocates keeping teams small enough to be fed with two pizzas.
As Amazon evolved, it shifted toward assigning a single accountable leader to large initiatives while preserving clear ownership and coordination.
Meta's reorganization appears to have done neither: teams remained very large without strong centralized ownership, contributing to confusion and declining morale.
The article warns that current trends toward eliminating middle management can backfire, particularly in knowledge work requiring creativity and close collaboration.
Leadership coach Beth Steinberg argues that managers with too many reports cannot effectively coach, develop, or support employees, leaving them only able to push work forward.
The overall lesson is that organizations seeking flatter structures should carefully balance efficiency with effective management, communication, and team size, especially in innovation-driven environments like AI.