Amazon memo on AI use in engineering draws scepticism over mixed messages and token-usage metrics

No CVEs, versions or technical vulnerabilities are discussed; this is a commentary on workplace policy for AI tooling. The article notes Amazon shut down KiroRank, an internal leaderboard ranking staff by token usage, after employees generated junk tasks ('tokenmaxxing'), and reports Meta had a similar 'Claudenomics' effort. Defenders and engineering leaders are advised by the memo's author to measure delivery outcomes, such as the 90th percentile of adoption teams compared with actual delivery, rather than raw token or tool activity.

For security analysts · Written by AI from 1 source · Wrong?