What's this about?
An engineering-leadership piece arguing that the largest available productivity gains come from how a team works together, not from tooling — and that AI is currently absorbing attention that would be better spent on the organisational problems.
Why does this argument keep returning?
Because the measurable costs in most engineering organisations are coordination costs: waiting for review, waiting for a decision, rework from a misunderstood requirement, work done twice by two teams. Faster typing does not touch any of them. It is the same observation as the "coordination headwind" piece below, and the same reason Amdahl's law is the most quoted result in performance work — speeding up the part that was never the bottleneck does not speed up the whole.
What's the community saying?
HN 448 points, 114 comments — high score, comparatively low comment count, which usually means broad agreement
- It landed the same week as No AI Fridays and Bug Blindness; HN's mood in this window is notably reflective rather than enthusiastic.
- The argument is not anti-AI, which is why it avoided the usual fight — it is about where the marginal hour of improvement effort goes.
- Worth reading against dotey's AI-native argument this week, which says the opposite: that redesigning processes around agents *is* the organisational change.