Amazon Warns Engineers Against Tokenmaxxing

Oct 27, 2026
Amazon says token use, AI adoption and generated code are flawed measures of engineering productivity.

Amazon is discouraging engineers from using AI tools without regard for useful output, a practice known as tokenmaxxing, according to CNBC. An internal memo says token use, AI adoption rates, code changes and code reviews are deeply flawed measures when used to assess teams or organizations.

The memo was circulated internally in September and published on Amazon's website on Tuesday. It says those metrics can still be considered with other data, but employees should focus on long term outcomes. Amazon has not imposed caps on employee token use, though it removed an internal leaderboard in May after workers used extra tokens to improve their rankings.

Amazon Web Services vice president Marc Brooker said the company also examines feature velocity, including the number of launches and successful changes made in production. The memo tells engineers not to use AI for tasks that do not benefit from current tools and says identifying where AI does not help is a valid finding.

We hope you enjoyed this article

Consider subscribing to one of our newsletters like Enterprise AI Brief or Daily AI Brief.

Also, consider following us on social media:

Enterprise AI AI Brief AI Brief (X)
Free newsletter

Enterprise AI Brief

Weekly report on AI business applications, enterprise software releases, automation tools, and industry implementations.

Guide
9 Voice to Text Apps

9 Voice to Text Apps

The best AI dictation apps for turning speech into clean text anywhere you type.

AI Token Webinar

Whitepaper

Tensordyne Napier: What If One Rack Could Do the Work of Nine?

Tensordyne

This Tensordyne whitepaper presents Napier, an inference-focused AI processor and rack-scale system based on the company’s TDN Math logarithmic number system. It examines infrastructure requirements for large mixture-of-experts and agentic models, compares major inference architecture approaches, and details the TDN AIP processor, TDN72 pod, TDN Link fabric, and Napier Ultra configuration. The paper reports simulation-based performance, cost, and accuracy-validation results, including Tensordyne’s projected comparison of one Napier rack with a nine-rack Nvidia Rubin plus Groq deployment; the chip is reported as taped out and in fabrication.

Read more

You may also like

Free, six days a week

Daily AI Brief: the AI news that matters, in your inbox.