Meta ends ‘tokenmaxxing’ by removing AI usage from employee evaluations

Meta is reportedly moving away from a practice known as ‘tokenmaxxing,’ where employees were encouraged to maximize their use of artificial intelligence tools. Previously, the company utilized AI adoption dashboards and token counts as part of performance evaluations, effectively rewarding heavy AI usage and creating competitive leaderboards among staff.

New internal guidance indicates that AI usage will no longer be a formal criterion for employee assessments. According to reports, Meta engineers have been informed that token counts will be removed from evaluations, and new directives suggest that desired work outcomes ‘can be supported by AI or other means.’

The shift follows earlier reports that external factors, such as Google throttling Meta’s AI tokens, forced the company to reconsider its approach to maximizing AI consumption. While the previous culture placed pressure on staff to use AI constantly, some employees have expressed relief regarding the updated performance review standards.