Reid Hoffman backs tracking AI token use, but warns it is not productivity

In short: Some companies are ranking employees by how much they use AI, and Reid Hoffman says the number can show adoption, but it should not be used as a direct measure of productivity.

What's going on

A workplace trend called “tokenmaxxing” is spreading in Silicon Valley. It means companies track and sometimes rank employees by how many AI “tokens” they use.

A token is a small piece of text or data that an AI system reads and writes when it answers a question (think of it like the “meter” on a taxi that counts the ride). Tokens are also how many AI services decide what to charge, so more tokens usually means higher cost.

The debate heated up after reports that Meta had an internal dashboard and leaderboard for token usage, and later shut it down after it leaked to the press. Critics argue that ranking people by tokens is like praising someone for spending the most money. They also warn it could encourage waste, like running tools just to push numbers higher.

LinkedIn co-founder and investor Reid Hoffman supported the basic idea in an interview at Semafor’s World Economy summit. He said token tracking can be a useful dashboard to see whether people across different roles are experimenting with AI. But he said it is not a perfect measure of productivity, and it needs context about what people are doing with the tokens.

What to watch

More companies are likely to try simple AI usage scoreboards because they are easy to measure. The key question is whether employers pair token counts with real outcomes, like time saved or better work, instead of treating token totals as a performance score.

Source: TechCrunch AI