- Project OT was initiated by Meta executives in January to reduce headcounts by up to 60% and implement AI for daily tasks, according to reports.
- Reports in March and April indicated that impending layoffs and AI training inputs negatively impacted employee morale.
- The first round of layoffs occurred in May, but Zuckerberg canceled a subsequent wave planned for November immediately after.
- In June, CTO Andrew Bosworth reported a 220% year-over-year increase in code changes, while user-facing features only increased by 36%.
- In July, Zuckerberg acknowledged that the development of AI agents had not progressed as anticipated.
- An October internal post detailed a pilot program aimed at removing middle management and utilizing agent-assisted analysis.
- Meta ultimately canceled Project OT before finalizing the number of layoffs, confirming that not all scenarios were pursued.
- Project OT aimed to replace many human employees with AI agents to perform daily tasks, as reported by internal documents.
- Internal posts indicated that AI agents executed large-scale, disruptive actions that humans typically would not, leading to a 40% increase in major technical problems.
- The time employees spent resolving these issues increased by as much as 70%.
Meta's Project OT, initiated by CEO Mark Zuckerberg, aimed to transform the company into an 'AI native' organization by potentially reducing headcount by 60%. The project explored using AI to handle daily tasks, overseen by small teams. However, it was scrapped before layoffs were finalized.12
Internal documents revealed that AI agents executed 'large-scale, disruptive actions' that humans typically would not, resulting in a 40% increase in major technical problems compared to the previous year. This surge in issues led to a 70% increase in employee time spent on problem resolution.1011
Despite the ambitious plans, Meta ultimately canceled Project OT, which had been in development since January. The company had intended to use savings from layoffs to reward high-performing employees, particularly those skilled in AI engineering. The restructuring included a pilot program that reformed engineering and research teams into smaller groups, aiming to enhance efficiency and reduce middle management.9
The decision to halt Project OT reflects ongoing challenges within Meta as it navigates the complexities of integrating AI into its operations while managing workforce dynamics.
“The plan, codenamed Project OT, was created in January and explored reducing headcounts by up to 60%, but Meta canceled it before determining layoffs. Internal posts also revealed that employee time spent resolving AI-caused issues increased by as much as 70%.”
Ars Technica
