WorkThe story, in brief

Microsoft CEO Satya Nadella admits he's a token-maxer, too: "It's addictive"

Microsoft's CEO admits what nobody else will: frontier models are addictive, even when they're wasteful.

Illustration of two anonymous hands arranging task cards around an amber tool on a shared desk.
People, judgement and the changing nature of work.AI illustration by KeyNews
The KeyNews take

Why it matters

Nadella's candid confession reveals a paradox at the heart of enterprise AI adoption—leaders know token-maxing is economically irrational, yet they can't stop doing it. This signals a gap between stated optimization strategies and actual deployment behavior that investors and operators need to watch.

The key facts

5 to know
  1. Nadella warns against 'token-maxing' (using frontier models for all tasks)

  2. Claims marginal productivity gains must justify token costs

  3. Admits to being a 'token-maxer' himself despite the warning

  4. Describes the behavior as 'addictive'

  5. Suggests frontier models are being wasted on everyday tasks

Go to the source

The Decoderthe-decoder.com

Publisher excerpt: Microsoft CEO Satya Nadella is warning against "token-maxing," throwing the most powerful AI models at every problem. Frontier models shouldn't be wasted on everyday tasks, he says. The marginal cost of productivity gains has to match the token cost. Still, Nadella admits, "I'm like a token-maxer…
Read original report
Back to today's editionMore work news

Keep reading

Related stories

More from Work