🔥 Meta Employees Reportedly Burning AI Tokens for Nothing — More Waste = Better Reviews

What happens when a company measures AI success by how many tokens employees consume?

Some Meta employees allegedly found the answer: run AI on meaningless tasks, loop pointless queries, and let the usage numbers soar — because higher consumption meant better performance reviews.

Reports reveal that Meta's internal AI usage hit 2 trillion tokens per day, but not all of that compute is productive. With KPIs tied to token consumption rather than actual outcomes, some workers gamed the system by burning through AI resources on empty tasks.

This is Goodhart's Law in action: "When a measure becomes a target, it ceases to be a good measure."

🎯 Why this matters:

• Meta spends billions on AI infrastructure — wasted tokens mean wasted money
• The problem isn't AI itself, but how organizations measure AI adoption
• Every company rushing to deploy AI faces this same trap
• Token volume ≠ business value, yet many orgs track exactly that

It's like judging a chef by how many dishes they wash instead of how the food tastes. The numbers look great on paper while real value goes up in smoke.

A cautionary tale for every organization measuring AI ROI: if you incentivize usage instead of outcomes, you'll get plenty of usage — and very little outcome.