The industry obsession with AI tools promising 2x, 5x, or 10x engineering productivity overlooks a more fundamental driver: organizational culture. Throughout a 13+ year engineering career, I've witnessed how toxic environments—where executives suggest AI might replace engineers—destroy psychological safety and halt collaboration. No AI tool can compensate for a culture where departments blame each other and communication breaks down.

Conway's Law explains why: systems mirror the communication structures of the organizations that build them. Bad culture produces bad products because people don't work together effectively. Good culture acts like human health—a prerequisite for everything else. Without it, AI investments yield diminishing returns.

Executives often panic when competitors report AI-driven productivity spikes, triggering FOMO-driven tool adoption. But tools layered onto dysfunctional culture amplify dysfunction. The teams actually pulling ahead invest in a "context layer" that gives AI agents understanding of system conventions and past decisions, not just raw data. Culture isn't soft—it's the hardest productivity lever, and the only one that compounds.