Amazon Warns Engineers Against Tokenmaxxing With AI
Left 67%
Center 0%
Right 33%
2 left · 0 center · 1 right
What happened
Amazon on Tuesday posted to its website an internal memo circulated in September that revises how the company defines the roles of software development engineers and managers in the AI era. The memo, which Business Insider attributed to vice president and distinguished engineer Terk Terkowitz, discouraged "tokenmaxxing"—using AI as much as possible with little regard for output—and described token usage, AI-tool adoption rates, and the volume of code changes or reviews as "deeply flawed" measures that can raise costs and complexity if treated as targets, saying it is "best to measure, and be measured by, long-term outcomes." Amazon has not instituted caps on employees' token usage; in May it shut down an internal leaderboard after discovering employees were tokenmaxxing to climb the ranks. The memo said the company is "certain" that AI will be "at the center of software engineering practice" but that "If your work does not benefit from current AI tooling, that is a finding, not a failing," and that an engineer who can explain where AI does not yet help "is demonstrating the judgement we are asking for." Marc Brooker, a vice president and distinguished engineer at Amazon Web Services, said Amazon evaluates the technology's effect on productivity through "feature velocity"—including launches and successful production changes—as a return-on-investment question of whether tokens and effort are spent on problems that matter.
Sourcing check · Partly on the record · 3 articles
1 of 3 articles cite a named source for the story. It was on the record from the first report.
What these terms mean and how they're counted
- On the record · names a source 1 of 3
- On the record means the article cites a source by name for the story's main claim: a person, or an institution speaking officially, such as a company statement or a named agency's spokesperson. The opposite is an unnamed source. The count is how many of the counted articles cite at least one such source; the names are everyone named across them, each once however the articles spell it. Here: Terk Terkowitz
- Primary evidence · cites a document or data 2 of 3
- Primary evidence means the article cites a record you could check yourself: a court filing or ruling, a dataset or report, a recording or transcript, an official statement or press release, market data. A quote is not a document. The count is how many articles cite one. Here: internal company memo, internal Amazon memo
- Credits another outlet 0 of 3
- The article says the story comes from another outlet's reporting (“first reported by Reuters”). Those articles relay the story; they are not independent confirmation. Here: Each article presents its own reporting.
- Unnamed sources 0 of 3
- The main claim is attributed to people who are not identified: “people familiar with the matter”, “officials said”. The count is how many articles do this.
- How it developed
- When the first report landed and whether it named anyone; when the first named source and the first document appeared. Here: First report Oct 27, 18:57 UTC by CNBC, naming its source. First document in the first report, from CNBC (internal company memo).
- Not counted 0
- Opinion, analysis and video pieces, and articles too short to assess. Every count above is out of the counted articles. An article republished under a new headline is one article.
This is about who is cited, not whether the story is true. Each source below says what it names. How we count
Omitted — what each side leaves out
Unpacked
Right-leaning coverage presents Amazon's memo as a cost-driven halt to padded AI use and skips the remaining pressure on light adopters. Left-leaning coverage, via Business Insider, reports that managers should inspect the bottom 10% of AI users and that teams not adopting are falling behind. Right-leaning coverage says the practice was "encouraged by tech giants" until huge usage bills, without that second instruction. That omission turns a change in how engineers are scored into a story about bills alone.
Right-leaning coverage says staff were "padding their AI usage just to look productive." Left-leaning coverage at CNBC says engineers should be "pickier about using AI."
What is Amazon's internal token bill, and will an engineer who documents that current AI tools do not help still get a clean performance review?
Bottom line
Amazon's memo is not just a ban on tokenmaxxing; Business Insider reports managers were also told to watch the bottom 10% of AI adopters, a detail Breitbart's account skips.
The receipts — 3 facts one side's coverage lacks
- Managers were told to pay particular attention to teams in the bottom 10% of AI adoption, and to the top 10%
- Amazon's most AI-forward frontier teams have shown stronger productivity and delivery
- Engineers were told they need to embrace the change and take time to learn, or they risk being left behind
The Left View
CNBC and Business Insider present the guidance as Amazon trying to capture real productivity gains from AI without letting usage metrics become the goal, against a backdrop of soaring industry AI costs and uncertain returns. They tie the memo to CEO Andy Jassy's rapid integration of AI and his prediction that it will "change the scope and speed" of work and could shrink Amazon's corporate workforce through automation while creating "other jobs." Business Insider foregrounds Amazon's "era of the Renaissance Developer," in which human judgment and machine intelligence combine, and reports that leaders are expected to use AI on production work themselves; that the most AI-forward "frontier teams" have shown stronger productivity and delivery; and that teams and individuals not adopting AI or increasing output metrics "are falling behind." Both outlets treat the earlier leaderboard as evidence that targeting tokens produced wasteful activity, and they highlight Brooker's yardstick as the alternative Amazon prefers to caps or cheaper-model switches.
The Right View
Breitbart casts the memo as a retreat by tech giants that had encouraged extreme AI use and are now "cutting back" because they "face huge usage bills," describing tokenmaxxing as employees "padding their AI usage just to look productive." It notes that other firms have capped tokens or shifted work to cheaper open-source models, "including Chinese alternatives," while Amazon has not imposed caps. The outlet pairs Jassy's comments on AI changing work and possibly shrinking the corporate workforce with a wider contest over who controls the technology, arguing the MAGA movement should take positions that "benefit humanity without handing control of our nation to the leftists of Silicon Valley or allowing the Chinese to take over the world."
Our Take (balanced)
Left-leaning coverage's strongest argument is that Amazon is correcting a measurement failure rather than cooling on AI: shutting the May leaderboard, treating unused tools as a finding rather than a failing, and judging ROI by launches and production changes all support a shift from activity theater to customer outcomes, while Business Insider's "frontier team" results are offered as evidence that serious adoption still pays. Right-leaning coverage's strongest argument is that the crackdown is a bill coming due: companies that ranked workers by tokens are now calling those same metrics deeply flawed after costs surged, and peers are already migrating work to cheaper, including Chinese, models. The central unresolved tension is whether Amazon can keep AI at the center of engineering practice—and treat non-adopters as falling behind—while honestly rewarding engineers who decline tools that do not help.
3 sources
More stories
The week's bottom lines, in your inbox
One email a week: the five stories that mattered and what they actually mean. Free.