A panel of senior engineering and research leaders from Etsy, Twilio, GitHub, Google, and DX debated key assumptions about AI's impact on software development. Topics covered include whether AI will reduce engineer headcount (consensus: no, demand will grow), whether AI accelerates technical debt (mixed views), whether most code will be AI-generated within five years (likely for new code), whether the future of engineering is about managing agents (nuanced yes), whether leaders should mandate AI adoption (no — adoption is inevitable), whether code review is now the bottleneck (real bottlenecks are planning and decision-making), and whether AI adoption barriers are really culture problems (yes, primarily human and organizational). A key throughline: AI reshapes the task mix but doesn't eliminate engineers, and the biggest risks lie in how leaders design roles and measure impact.
Table of contents
Will AI mean fewer engineers?Is AI accelerating technical debt?Will most code be AI-generated within five years?Is the future of engineering about managing agents?Do leaders need to mandate AI adoption?Is code review now the bottleneck?Is risk the only thing holding back adoption?Are AI adoption problems really culture problems?Sort: