Dave Rupert has a theory: when organizations prioritize speed, communication dies first. In his essay, Rupert argues that "moving fast" shortcuts the slow, necessary work of consensus-building. Teams stop consulting subject matter experts. Cross-team collaboration suffers. People start building in different directions, setting up what he calls "messy merge conflicts" down the road. Design systems and codebases fragment too. When deadlines loom, developers detach Figma components and duplicate folders rather than fix the underlying system. Technical debt piles up, hidden until someone else inherits the mess.
Rupert's sharpest point is about AI. He calls LLMs the ultimate "Don't talk to my coworkers" toolchain. Why seek out an expert who might push back when you can get instant validation from a machine that always says yes? Hacker News readers saw the same thing. They pointed out that moving fast without coordination often produces broken systems and features nobody asked for. The friction of human disagreement serves a purpose. Skipping it doesn't make you faster. It makes future conversations harder because you've already sunk costs into a direction no one agreed on.
There's a subtler problem too. When developers solve problems through AI chat interfaces, the reasoning disappears. No paper trail. No shared context in the codebase. The solution lands in your repo, but the rationale stays trapped in a private LLM session. Six months later, someone else has to reverse-engineer why that code exists at all. Rupert wants engineering management to stop treating their job as pushing tickets across a board. The real work is getting everyone rowing the same direction. That requires talking. It requires slowing down.