May 11 · 3 min read · The CLI could create a credit note. That was the bad news. The umbrella documents compose command was built to make all document types reachable through one stable surface. Pass a type, an entity, and a JSON body. The server looks up the per-type Zod...
Join discussionMay 11 · 17 min read · I've seen teams burn weeks untangling API messes that could have been avoided with a couple of days of upfront thinking. Endpoints named inconsistently, versioning added as an afterthought, error resp
Join discussion
May 7 · 3 min read · Understanding the gap between actual performance and perceived latency In previous parts, we explored how backend systems behave under load and how design decisions impact performance. However, not all performance issues come from slow systems. In ma...
Join discussionMay 2 · 12 min read · Every ASP.NET Core team eventually faces the same question: how should our API communicate with its consumers? Not just when something goes wrong — but in every response, success or failure. The decis
SRobert and 1 more commented
Apr 29 · 5 min read · Enterprise contract management systems process millions of documents annually while maintaining strict security, compliance, and performance requirements. Architecting such platforms requires careful consideration of data models, integration patterns...
Join discussionApr 26 · 3 min read · There's a quiet revolution happening in how developers think about AI agent security, and it has nothing to do with the flashy debates about alignment or consciousness. The real engineering challenge is far more mundane: how do you give an AI agent a...
Join discussion
Apr 26 · 3 min read · A conversation on V2EX, China's vibrant developer forum, surfaced something interesting this week. A developer was trying to use GitHub Copilot's GPT-5.4 model—routed through Copilot's infrastructure—inside Anthropic's Claude Code environment. The qu...
Join discussion
Apr 25 · 4 min read · The developer community has entered an era of AI tool stacking, where Claude Code calls GPT models, Copilot wraps Claude APIs, and developers casually route inference through services that were never designed to be intermediaries. A recent discussion...
Join discussion
