🔍 Read the full analysis: A Practical Look At AI Outcomes Through Amazon’s Investment Strategy on ThorstenMeyerAI.com
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Amazon published engineering guidance on October 27 saying software engineers should be judged by outcomes, not token use or AI-tool adoption. The memo comes as Amazon’s in-house Nova model effort has been scaled back, according to reporting, while its Anthropic investment has risen sharply in value on paper. Amazon’s third-quarter results are due October 29.
Amazon’s engineering leadership told software engineers on October 27 to focus on the effects of their work, not how many tokens they use or how often they adopt AI tools. The guidance matters beyond employee evaluations: it arrives as the company’s Nova models have been scaled back, according to Business Insider, while its investment in Anthropic is valued at substantially more than its original cost on paper.
The memo, published by Amazon and attributed to Terk Terkowitz, a vice president and distinguished engineer in its stores business, introduces a revised role guideline for software engineers. Written by senior engineers, it avoids prescribing particular AI tools or usage levels. Instead, it says engineers’ impact on customers, the business and their teams should remain the basis for evaluating their work.
The guidance calls token usage, AI adoption, code changes and reviews “deeply flawed metrics” when used as targets rather than measurements. Such goals can raise costs and add complexity, the memo says, citing tokenmaxxing as inconsistent with frugality and measuring lines of code as inconsistent with simplifying systems. It also says these indicators currently “correlate well” with important outputs, preserving a distinction between using data to monitor work and using it to score employees.
Amazon recommends that managers examine both low and high adoption levels to understand where AI helps and where it does not. The memo says that work which does not benefit from current AI tools is “a finding, not a failing,” and asks engineers to explain their observations. Leaders are also told to use the tools on their own production work and compare tasks where AI helped with those where it did not.
Six months at Amazon
From token leaderboards to an outcomes memo
Amazon spent 2026 learning the same lesson twice: once with how its engineers use AI, once with its own AI models.
- May 2026Token leaderboards backfireFinancial Times (via The Decoder): staff give an internal agent tool pointless tasks to inflate token counts. Reported, anonymous sources.
- Late May 2026KiroRank “deprecated”Amazon spokesperson to Business Insider. SVP Dave Treadwell: “Please don’t use AI just for the sake of using AI.”
- 24 Jul 2026San Francisco AGI Lab site closesConfirmed by Amazon to GeekWire. Frontier model research continues under Pieter Abbeel.
- 28 Jul 2026Most flagship Nova models wound downBusiness Insider, people familiar: Premier, Omni, Reel, Canvas to “keep the lights on”. Amazon: models “remain one of the most important things we’re working on”.
- 14 / 30 Sep 2026Nova Sonic v1, Nova Reel, Nova Canvas reach end of lifeAWS Bedrock model lifecycle page. Requests to them now fail.
- 26 Oct 2026Anthropic files its S-1Amazon notes: $6.2 billion principal, 12% ownership cap that Amazon can waive. Anthropic commits at least $110 billion to Amazon through April 2036.
- 27 Oct 2026The memo: outcomes, not tokenmaxxingUsage and throughput metrics are “deeply flawed” as targets. “Tokenmaxxing is at odds with frugality.”
- 29 Oct 2026Amazon’s third-quarter resultsAnnounced by Amazon. Not yet published when this article went live.
Sources: Amazon engineering memo (27 Oct 2026); Business Insider (29 May and 28 Jul 2026); GeekWire (24 Jul 2026); The Decoder (12 May 2026, summarising the Financial Times); AWS Bedrock model lifecycle (read 28 Oct 2026); Anthropic S-1 (26 Oct 2026).
Powered by Thorsten Meyer AI
Amazon’s AI Bets Face Outcome Tests
The memo turns a broader debate about AI productivity into an internal management rule: activity is not the same as impact. Counting tokens or tool adoption may show that employees are using AI, but it does not establish that customers receive better products, work becomes faster, or costs fall. Treating those counts as targets can reward unnecessary use, the memo warns.
Amazon’s investment record offers a separate, imperfect illustration of why outcomes matter. The source material says the company invested about $18 billion in Anthropic, and its June 30 filing valued its holdings at about $190 billion. That figure is a reported paper valuation, not a cash return or a measure of Anthropic’s operational contribution to Amazon. The source also says gains on the stake in the second quarter exceeded Amazon’s operating income for that period; the comparison concerns investment gains and operating income, not equivalent measures of ongoing business performance.
For workers and other companies, the practical implication is that AI use alone is a weak proxy for value. Amazon’s position is not that usage data should disappear, but that managers must connect it to results and investigate cases where the tools fail to help.
Products Worth Considering
Amazon Kindle Paperwhite (12th Generation, 2024), 16 GB, 7-inch Display, Jade
Amazon Kindle Scribe Colorsoft 64GB (1st Generation)
Amazon Kindle Paperwhite Signature Edition (12th Generation, 2024), 32 GB, Metallic Black
As an affiliate, we earn on qualifying purchases.
From Usage Leaderboards to Guidance
The memo follows earlier scrutiny of internal AI-use tracking. In May, the Financial Times reported that Amazon employees gave an internal agent tool, MeshClaw, unnecessary tasks to raise their token counts. Those accounts were anonymous and remain reported allegations, not a company-confirmed account of widespread behavior. The same reporting described a weekly AI-use target and internal usage leaderboards.
Amazon did confirm to Business Insider at the end of May that a leaderboard known as KiroRank had been “deprecated.” The company described it as an informal tracker created by a group of employees and said it was not intended to promote AI use for its own sake. Senior vice president Dave Treadwell also told staff, “Please don’t use AI just for the sake of using AI.” The October guidance formalizes a related principle in the company’s engineering role expectations.
Amazon’s model portfolio provides another part of the backdrop. Business Insider reported in July that most flagship Nova models had been moved to a “keep the lights on” status. Amazon’s lifecycle information lists Nova Canvas, Nova Reel and the first Nova Sonic as retired. These developments do not establish that every Nova product or model has ended, nor do they by themselves explain the commercial reasons for the changes.
““Tokenmaxxing is at odds with frugality.””
— Amazon engineering memo
What the Scorecard Cannot Show
The memo does not publish measured productivity gains, cost savings or customer outcomes attributable to AI use across Amazon’s engineering teams. It says adoption indicators correlate with results today, but gives no figures, time period or method for that relationship. It is also not clear how individual managers will weigh outcome measures against usage data in performance reviews.
The investment figures need similar care. The reported $190 billion valuation is the value Amazon’s filing assigned to its Anthropic holdings as of June 30, according to the source material; it is not a realized profit. The reasons for the Nova changes, their effect on Amazon’s AI services, and the extent to which Amazon relies on Anthropic models remain unclear from the material provided. Amazon had not yet reported its third-quarter results when the source was written.
Third-Quarter Results Due October 29
Amazon is scheduled to report its third-quarter results on October 29. Those figures may offer further information about the company’s financial performance, but they will not by themselves show whether engineers’ AI use has improved customer outcomes or reduced costs. The memo does not set a date for follow-up measurements or describe a companywide audit of the recommended practices.
The immediate test is how teams apply the guidance: whether managers use adoption data to identify barriers, check whether heavy use produces better work, and document tasks where AI adds no value. Until Amazon publishes results on those questions, the memo states a measurement principle rather than demonstrating its effects.
Key Questions
What did Amazon tell its software engineers?
Amazon’s October 27 guidance says engineers should be judged by their impact on customers, the business and their teams, rather than by token use or AI-tool adoption alone.
Is Amazon telling employees to stop using AI?
No. The memo says AI-adoption indicators currently correlate with important outputs. It argues against turning those indicators into targets and asks teams to examine whether AI use improves results.
What happened to Amazon’s Nova models?
Business Insider reported that most flagship Nova models were moved to a “keep the lights on” status. Amazon’s lifecycle information lists Nova Canvas, Nova Reel and the first Nova Sonic as retired. The available information does not say that every Nova model or product has ended.
How much is Amazon’s Anthropic investment worth?
The source material says Amazon invested about $18 billion and that a June 30 filing valued its holdings at about $190 billion. That is a paper valuation, not a realized cash return.
When is Amazon’s next results report?
Amazon’s third-quarter results are scheduled for October 29, 2026, two days after the engineering guidance was published.
Source: ThorstenMeyerAI.com
