We trust open-source code we've never read because of a stack of primitives the ecosystem agreed on. AI-generated code is a new kind of third-party code, and most of those primitives don't exist for it yet. A first cut at the list.
Stack Overflow helped train the AI models that are now making it irrelevant. As developers solve problems privately with AI assistants, the public knowledge commons is quietly starving. This is an underrated problem.
The modern equivalent of answering Stack Overflow questions is contributing fixes to open source issues. AI is what finally makes this approachable at scale.
AI agents can generate 10x more code, but can humans review it? Exploring why code review is the new bottleneck and a controversial solution - treating AI-generated code as third-party dependencies.