According to the company's internal statistics, about 93% of codebase changes are already created using Agentic AI. The practice of "tokenmaxxing" (where employees artificially generated gigantic volumes of code via AI just for reporting) led to bloating technical debt and a drop in architectural quality. Meta's new instructions refocus on utilitarian business metrics: Time-to-Market, task complexity, and product stability. For the B2B industry, this is a powerful marker of maturity: corporations are stopping the game of "who can burn the most tokens" and are beginning to demand real financial profitability from algorithms.
Source: Meta / The Information
ManagementMetaAgentic AIB2BMacroeconomics