Uber executive questions link between AI token use and productivity gains

Uber's chief operating officer said in an interview released last week that rising AI token usage has not produced clear productivity gains. The executive stated the direct link between token consumption and output remains difficult to establish. The comments, which received more than two million views on X, noted that while more work may be completed, connecting specific usage statistics to concrete results such as additional consumer features is challenging.

AI tokens represent the basic units processed by AI systems, with each token roughly equivalent to three-quarters of a word.

spending patterns Several large companies have adopted internal AI programs that track employee usage. Meta designates some staff as AI builders working in specialized teams, while Disney and JPMorgan monitor employee AI activity. Visa has reported monthly token consumption approaching two trillion.

Uber exhausted its annual AI budget within the first four months of the year, according to separate reporting cited in the article. Multiple technology professionals have posted that a significant portion of internal token spend yields no measurable return.

Industry responses and guidance Google's chief executive said at the company's developer conference last week that chief information officers have expressed concern over budget overruns. The executive added that the issue is expected to intensify through the remainder of the year.

An engineering intelligence firm report found that the top 10 percent of users of one AI coding tool consumed roughly ten times the median number of tokens yet produced only twice the output. The report recommended tying AI costs to concrete metrics such as code changes submitted rather than raw token counts.