Jensen Huang has introduced a new test for engineers at Nvidia, one that assesses whether an engineer is worth keeping based on their token consumption. The test requires the $500,000 engineering team at Nvidia to come in under half of their annual AI token consumption if they meet certain salary standards. According to an interview with Jensen Huang, Nvidia chief executive, this test aims to identify engineers who are not contributing significantly enough to the company's success and potentially be let go.
Huang has stated that he will only keep employees whose AI token consumption is below 50% of their annual salary. This means that if an engineer consumes tokens at or above 100%, it would indicate they have a high level of contribution to the team, making them worth keeping. The goal is to maximize budget potential without sacrificing team success. By attaching a token budget to employee salaries and performance, Nvidia aims to encourage engineers to optimize their usage of AI resources.
The test has been implemented as part of Nvidia's effort to maintain a lean team structure while still delivering high-quality products and services. Engineers who are found not to be contributing enough are expected to work towards reducing their token consumption or exploring other roles within the company. By setting clear expectations around salary and AI token usage, Nvidia can better manage its workforce and ensure that engineers are utilized effectively without sacrificing overall team success.