Meta CTO Bosworth Admits 'Atrocious' AI Reorg, Vows Changes

Meta's chief technology officer, Andrew Bosworth, acknowledged in an internal post that the company's restructuring of its Applied AI unit was 'atrocious' and pledged to restore employee trust through better communication, career growth, and scaled-back manager spans. The division, formed in March with roughly 6,500 staffers, had been criticized by workers for menial assignments—one employee likened it to 'a gulag.' Bosworth's memo, seen by WIRED, responded to widespread dissatisfaction reported last week. He promised to cap managers at 20 direct reports, limit the frequency of manager switches during reorganizations, and allow employees to use internal 'AI coaching' tools. Bosworth also acknowledged the leadership lost sight of individual career trajectories amid strategic pivots like the rush to compete in AI coding tools.

In a related post, Maher Saba, VP leading the Applied AI team, told employees they could now seek other roles within Meta—reversing the initial draft policy that forced people onto the team. Saba emphasized the group's focus on enhancing Meta's frontier AI models' coding and agentic capabilities, and said traditional roadmaps no longer apply because 'the work evolves so dynamically.' Despite the mea culpa, Bosworth maintained that the speed-driven drafting was necessary and warned that some projects may require 'sacrifice.' The unrest is part of broader morale issues at Meta following mass layoffs and surveillance concerns, with CEO Mark Zuckerberg also posting recent messages to address employee sentiment.