Meta Pushes Its New AI Agent on Employees—but Eases Off on Tokenmaxxing (www.wired.com)

Meta has announced a significant shift in its employee performance evaluation approach by discontinuing the controversial "tokenmaxxing" system, which previously rewarded workers based on their AI tool usage. The new guidelines now focus on evaluating employee impact without emphasizing AI usage metrics or designations like “AI Native.” This change aims to alleviate pressure on employees to use AI tools inappropriately just to boost performance scores, thus improving morale and rebuilding trust after previous controversies around employee surveillance and AI implementation. Simultaneously, Meta is testing its new AI tool, Hatch, designed to autonomously perform tasks across various applications, which has generated interest but also privacy concerns among employees. While the tool shows promise for enhancing productivity—some workers have reported using it to manage personal tasks—there is apprehension about the environmental impact of its high token consumption. Although employees are reassured that their usage won’t directly affect their evaluations, the pressure to demonstrate effective usage remains, raising concerns over potential further layoffs if Hatch's implementation leads to increased efficiency at Meta.