Meta is formally ending a performance incentive program that led to a practice known as "tokenmaxxing" among its workforce. In an internal announcement, the social media giant informed employees that their performance evaluations will no longer depend on how much they use AI tools, shifting the focus back to overall impact.
The End of Tokenmaxxing
Nearly a year ago, Meta began grading workers on their "AI-driven impact," which effectively measured how much they utilized chatbots and agents in their daily routines. This led to employees receiving labels such as "AI Native" or "AI Enabled." The pressure to perform according to these metrics resulted in some workers repeatedly prompting AI tools simply to boost their internal usage statistics. One employee even created an internal leaderboard to track token consumption, labeling high-usage peers as "Token Legends."
The updated guidance replaces these rigid criteria with looser wording, stating that professional outcomes can be supported by "AI or other means." This shift follows a lawsuit from approximately two dozen former employees who allege that the previous metrics unfairly penalized those on health or family leave who could not accumulate usage data.
Hatch: The Next Frontier
While formal pressure is easing, Meta is encouraging employees to experiment with "Hatch," a new agentic AI tool. Similar to the viral "OpenClaw," Hatch can autonomously perform tasks on a computer, such as browsing the web and operating various applications. While some employees report using Hatch for personal tasks like booking appointments, others remain hesitant due to privacy concerns. These worries stem from a previously paused Meta project that tracked employee keystrokes to gather training data for AI systems.
Productivity and Job Security
The introduction of more powerful tools like Hatch brings mixed emotions to the Meta workforce. While the tool is described as effective, there are lingering fears regarding job security. Although CEO Mark Zuckerberg has stated that mass layoffs are not expected for the remainder of the year, some employees worry that significant productivity gains from AI could eventually lead to further workforce reductions. Reports suggest that a recent round of planned layoffs was called off due to software bugs and slower-than-expected progress in AI agent development.
