Meta has recently launched and then quickly dismantled an internal AI leaderboard called “Claudeonomics” designed to promote the use of generative AI among employees. This system tracked employees’ token consumption—units measuring input and output in AI language models—to monitor and gamify AI usage, assigning badges to top users. Despite Meta’s intention to foster AI integration into daily tasks, concerns arose regarding the sharing of internal usage data, leading to the leaderboard’s shutdown shortly after its inauguration. In just 30 days, Meta employees reportedly consumed 60 trillion tokens, with the top user accounting for 281 billion. This initiative reflects a broader industry trend toward monitoring AI engagement as a productivity metric, a practice termed “tokenmaxxing.” Industry leaders, including Nvidia’s CEO Jensen Huang, advocate for making AI usage a fundamental part of employee compensation. As AI tools become essential to workflows, a shift in evaluation standards is anticipated in Silicon Valley.
Source link
Share
Read more