Code review has changed shape every time its constraints moved, and agentic code generation has moved them again. We hand line-by-line review to agents and put human review where it matters most: running the change and building a shared mental model of the system.
Read this essay →
We cut a third of the engineering team in April and reorganized what was left around agentic development. Four months later, features shipped are up 120% and the DORA metrics are flat or better. Here is what moved, and where the costs showed up instead.
Reshaping our organization around agentic software production sent us back to our hiring rubric early, and technical ability lost its standalone score. Here are the six criteria we landed on, and the thinking behind each.
I shipped code to production as a designer. Some unexpected things happened when I coded with Claude and I learned why scope creep is so prevalent. In the end it made me think about new design process ideas I want to try implementing.
Subscribe via RSS, Atom, or JSON Feed — or see all options.