Amazon Finds 'Tokenmaxxing' Lacks Productivity Gains
Ancient Slashdot reader Gilmoure shares a report from Tom's Hardware, adding: "Huh. Looks like consuming one's own slop does not turn the wheels of industry." From the report: Amazon wants its engineers to quit chasing token counts, warning that heavy AI usage does not automatically mean better work. In updated engineering guidelines published on October 27, the company laid out how developers…
Amazon is warning its engineers against focusing solely on the number of artificial intelligence tokens used, as heavy AI usage does not necessarily lead to better results. The tech giant published updated engineering guidelines on October 27, emphasizing that metrics like AI adoption rates, token usage, code changes, and code reviews are misleading when not tied to actual outcomes.
Amazon argues that metrics disconnected from real results encourage wasteful behavior, contradicting its values of frugality and simplicity, and may increase compute costs and software complexity. Instead, the company advocates for measuring long-term results as the primary yardstick. While input metrics are still valuable when analyzed carefully alongside other indicators, Amazon suggests comparing teams that use AI the least and the most to assess whether adoption truly enhances delivery.
The company also rejects the notion that every task requires an AI tool. If existing tools do not improve work, Amazon considers that a finding, not a failure, and engineers who can demonstrate where current tools fall short are showing good judgment rather than being reluctant to adopt AI.
Written by urgent.news from Slashdot's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
