https://gduv.club/articles How AI systems get put together in practice: context engineering, agent architecture, MCP versus plain scripts, and the token arithmetic behind each choice. en https://gduv.club/articles/mcp-vs-direct-apis https://gduv.club/articles/mcp-vs-direct-apis MCP is great for one-off jobs. For anything I run in volume it's slow, it burns tokens, and it makes my agent retype data it already has. Here's what I use instead: direct API calls in scripts, a skill that documents them, and a proxy so nobody keeps keys on their laptop. Wed, 16 Sep 2026 00:00:00 GMT https://gduv.club/articles/harness-agnostic-workspaces https://gduv.club/articles/harness-agnostic-workspaces I use four different AI coding tools in a normal week. Three practices keep my repositories working with all of them, and one problem I still haven't solved. Wed, 12 Aug 2026 00:00:00 GMT https://gduv.club/articles/small-models-around-big-ones https://gduv.club/articles/small-models-around-big-ones Letting a big model orchestrate smaller ones is now common. I think the bigger win is putting small models around it instead: picking its context, filtering what its tools return, and compressing results while they're still fresh. Tue, 15 Sep 2026 00:00:00 GMT