Meta's experiment in building an "AI-native" workforce has hit a wall—for now. Reuters reports the company quietly mapped out a sweeping restructuring called "Project OT" that envisioned shrinking some teams by up to 60%, swapping large groups of staff for small "pods" of workers supported by AI "agents." The overhaul was to roll out in two rounds of cuts, in May and November, and leaned on tracking US employees' mouse clicks and keystrokes to train the software that might one day replace portions of their jobs. Meta confirms the project existed, but says the 60% figure applied only to certain teams in scenario planning, not the company as a whole.
Zuckerberg went ahead with a roughly 10% layoff in May, then abruptly scrapped planning for the second wave amid mounting pushback and underwhelming AI results. At that point, Meta workers were in "open revolt," Reuters reports. Internal data showed code output exploding but user-facing improvements rising far more modestly—and technical and security incidents jumping 40%, with the time workers spent on those problems surging 70%. According to internal posts, the AI agents that were supposed to replace workers made "large-scale, disruptive actions that humans are unlikely to execute," per Ars Technica.
Employee sentiment plunged, labor organizing gained steam, and executives tried to steady morale with perk upgrades and a softer public line: Meta now touts itself as "betting on people," even as it pours an estimated $130 billion into AI infrastructure and keeps some of the small AI-driven teams in place. At a company town hall last month, Zuckerberg admitted that he had overestimated how well the AI would perform, saying Meta's investment in AI agents hasn't "come to fruition yet."
,
uBlock 
,
and
Ghostery 
)
