Meta Pushes Its New AI Agent on Employees—but Eases Off on Tokenmaxxing
wired ·
Meta is ending the practice of tying employee performance evaluations to their usage of internal AI tools, according to an internal announcement. Previously, workers were graded on 'AI-driven impact,' which some employees felt pressured them to maximize AI token consumption, a behavior dubbed 'tokenmaxxing.' The new guidance focuses on employees' overall impact and no longer uses metrics like 'AI Native' designations or token counts. Meanwhile, employees are testing a new agentic AI tool called Hatch, which can autonomously browse the web and operate applications. Some employees are hesitant to connect Hatch to personal accounts due to privacy and error concerns, partly because Meta previously paused a project that tracked employee activity. Others have embraced Hatch for personal tasks, though they worry about computing costs and potential job cuts if productivity rises. Meta says the updates emphasize evaluating contributions, and it declined to comment on Hatch testing.
wired