Zuckerberg Admits Mistakes in Meta AI Workforce Shift

Meta CEO Mark Zuckerberg told employees in an internal memo, seen by Reuters, that the company "made mistakes" in its AI transformation. The May restructuring laid off roughly 8,000 employees - about 10% of Meta's global workforce - and transferred 7,000 more into new AI-related initiatives. "Given the complexity of these changes, we've made mistakes and will almost certainly make more," Zuckerberg wrote, adding he is "focused on providing as much stability as possible" in further organizational changes. "I don't want to overpromise because the world is changing in ways that are out of our control," he said. He reiterated that Meta does not expect more company-wide layoffs in 2026. The memo states Meta will try to find new internal roles for reassigned staff: "By creating important new roles for people, this also allowed us to shrink the size of teams knowing that if we make mistakes in some places, then we could transfer some people back," Zuckerberg said, per Reuters. Meta declined to comment.

The memo arrives as Meta's Applied AI Engineering unit - roughly 6,500 engineers and product managers - is facing an acute morale crisis. TechCrunch reported on June 12, citing a Wired investigation, that the unit is "on the verge of revolt." Engineers describe being given a stark choice: join or quit. Many call themselves "draftees." Their assigned work centers on generating puzzles and coding problems to train AI models. "It's literally the gulag," one employee told Wired. "Most people find the work soul-crushing," said another. The unit's original structure placed up to 50 individual contributors under a single manager - a flat ratio that left employees without meaningful management support. More than 1,600 Meta employees company-wide separately signed a petition protesting a program that monitors their clicks and keystrokes for AI training data. Meta CPO Chris Cox described the past few months as "brutal" on an employee call this week, per TechCrunch. The Applied AI unit is led by Maher Saba, a 12-year Meta veteran formerly at Reality Labs, reporting to CTO Andrew Bosworth.

Zuckerberg's memo outlines several corrective steps: higher budgets for offsites and corporate events, a large-scale company-wide hackathon in July to foster cross-team collaboration, restoration of assigned desks in many offices by year-end, and a planned rollback of the broadened manager-span ratios, per Reuters.

In April, Meta raised its annual capex forecast to $125B-$145B, reflecting the scale of its AI infrastructure commitment. Zuckerberg explained the logic behind using internal engineers for data labeling in a leaked April audio recording: Alexandr Wang - who sold data-labeling startup Scale AI to Meta for $14.3 billion before taking the chief AI officer role heading Meta Superintelligence Labs (per TechCrunch) - knows the data-labeling world well, and Meta believes its average employee has "significantly higher" intelligence than outside contractors. An internal announcement reviewed by Business Insider found that Meta's AI models still lacked the knowledge to outperform humans at technical coding tasks. Combined with the 2022-2023 round of over 21,000 cuts, Meta has now eliminated roughly 30,000 positions over approximately four years.

• •Whether the July hackathon and increased team-building investment reduce attrition in the Applied AI unit.
• •Progress on finding internal roles for reassigned employees, and whether the transfer pathway reduces turnover.
• •Future reporting on model-quality improvements tied to the internal data-labeling program that motivated the restructuring.