Days after Meta shut down its internal “tokenmaxxing” dashboard following news that its AI leaderboards had been leaked to the press, LinkedIn co-founder and venture capitalist Reid Hoffman has voiced support for the concept that has recently taken Silicon Valley by storm.
AI tokens are small chunks of data that an AI model processes as it tries to understand a prompt and generate a response. It is also the unit used to measure AI usage and determine the cost of AI services.
As a result, many companies are starting to internally track which employees are using the most tokens as a way to understand which employees are more actively adopting AI tools. They call this concept “tokenmaxxing”. “maxxing” is Gen Z lingo for optimizing something, and you may have heard it in other slang terms like “looksmaxxing” and “sleepmaxxing.”
But engineers at tech companies are debating whether the metric is a viable measure of workplace productivity because it’s akin to ranking people based on whether they spend more money than others.
In an interview aired this week on Semaphore’s World Economic Summit, Hoffman said he offers advice to companies implementing AI and has a positive view of the practice. Although he didn’t mention this metric when talking about Gen Z, he said it’s a good idea to track employee token spending.
“We need people from all kinds of departments to really engage and experiment with[AI],” Hoffman said at the event. “This is one of the great things to look at in a dashboard. It doesn’t mean it’s a perfect example of productivity, but… how much token usage are people actually using?”
He went on to explain that while some people may be using large amounts of tokens, they may be using them in a more random or exploratory manner, so tracking ‘token maximization’ practices needs to be combined with understanding what people are doing with their tokens.
“Some of the experiments are going to fail, and that’s fine, but it’s in the loop and basically requires different people to use it collaboratively and at the same time,” Hoffman added.
Hoffman also shared other advice for companies looking to develop an AI strategy, suggesting they should embed AI throughout their organizations. I also suggested checking in regularly and sharing what’s working with others.
“Essentially, you need to check in weekly. You don’t need everyone to be together all the time, but you need to check in as a group about, “Individually and as a group, what have we tried new or learned this week to use AI to improve the productivity of our company?” Because some of the things you find there are really surprising,” Hoffman said.
