Amazon issues new AI guidelines for software engineers

Amazon has released updated engineering guidelines that prioritize long-term business outcomes and customer impact over raw artificial intelligence usage metrics. In an internal memo, the company warned against “tokenmaxxing,” a practice where employees deliberately inflate AI token consumption, tool adoption rates, or code volume to appear more productive.

Vice President and Distinguished Engineer Terk Terkowitz described the current shift as the era of the “Renaissance Developer,” where engineers combine machine intelligence with human judgment and domain expertise. The company noted that AI is transforming software development more fundamentally than previous shifts like cloud computing or microservices.

While Amazon encourages developers to integrate AI into their daily workflows to avoid falling behind, leadership cautioned that using input metrics as performance targets can lead to increased computational costs and unnecessary system complexity. Instead, the company intends to measure success through measurable business value and the quality of delivered products.