Losing your best employees doesn't always happen through direct action. Simply de-soul the job. Meta learned this the hard way, and the receipts are ugly. ## What actually happened
From March to June 2026, Meta relocated approximately 6,500 engineers and product managers to a new data labeling unit, Applied AI, representing 30% to 50% of staff on some core product, infrastructure, and security teams. They were not junior employees, but rather the individuals responsible for constructing and operating Meta's infrastructure. Their new role? Creating coding puzzles, Evaluating AI-generated GitHub repos, and Doing RLHF grunt labor. Management's message was clear. "Transfers are not voluntary."
The context nobody says out loud
In 2025, Meta acquired a 49% share of data-labeling firm Scale AI for $14.3 billion, and its founder, Alexandr Wang, was appointed Chief AI Officer. There was a logic behind the reassignment. A decision was made that the high-cost senior engineers would be operating the low-cost labeling machine. Meta posted a net profit of $26.8 billion in the same quarter it laid off 8,000 people and drafted 6,500 more into labeling. It was not about surviving. It was about making a decision. 💸
Layoffs aren't what breaks morale
Everyone braces for layoffs. Instead, they bleed out slowly as the survivors look over their shoulders, lose focus, and underperform. The silent killer of a team is how you reward the ones who remain. Engineers who spoke to WIRED described the labeling work as "soul-crushing." One said the Applied AI unit had become a "gulag" where, thanks to the lack of meaningful work, "you have zero purpose in life all of a sudden."
They began to refer to themselves as "draftees". You've already lost your top talent if they describe their current role as "conscription", they just haven't put in their two weeks' notice. Leadership was also aware. CTO Andrew Bosworth mentioned during an internal "Tuesdays with Boz" meeting on June 2, 2026, that the AI reorg had been "atrocious" and that morale was "maybe not the worst it's ever been in 20 years here, but it's probably up there"
The quiet resignation wave
The headlines didn't capture what really happened after the forced reassignments. As documented by Gergely Orosz of The Pragmatic Engineer, top infrastructure and AI engineers who were reassigned started actively interviewing. They left for Anthropic, OpenAI, and Google. The good ones walked. Nobody fired them. After that, panic set in. Meta changed its longstanding policy of not engaging with departing staff. They began providing senior IC6+ engineers who were attempting to leave with one-time retainer equity grants, frequently $1M+ vesting over 3 to 4 years. Several engineers responded with a negative answer. Consider this. No amount of money could erase the experience of being conscripted into a job that you feel has no purpose. ## What I take from this
I am in charge of a small startup. We are fifteen people, with a limited budget, and surprisingly enough, there are no forced labor camps involved. I may not have the highest bid, but I can be the reason you don't have to waste your time and resent your effort. The lesson here could not be more simple:
→ People stay for the work, not just the paycheck. → Once you strip the purpose out, money becomes a bribe, not a reason. → The talent you most want to keep has the most options to leave. → How you treat survivors after a cut speaks louder than the cut itself. Meta could afford $14.3 billion for Scale AI and $1M retention grants. What it couldn't replace, of course, was the feeling that the job was a meaningful one. It is easy to retain morale with minimal costs, but very difficult to restore it once it is lost. My question to you is: Have you ever quit a job not because of the pay or a layoff, but rather because the work slowly lost its meaning for you?
