There's so much confusion in the agent space – frameworks, patterns, "agentic" this and that. I kept finding myself unable to clearly explain what was actually happening in these systems, or where to look when something broke.
So I took a different approach: reverse-indexed from typical "intelligent" agent behaviors and asked "what code patterns actually implement this?"
The result is 10 elements that map behaviors like "remembers what I said" or "works while I sleep" to concrete code – conversation arrays, cron jobs, retrieval pipelines, etc.
It's a conceptual map, not a how-to guide. Hopefully useful for anyone trying to make sense of this space.
Author here. Built this out of frustration.
There's so much confusion in the agent space – frameworks, patterns, "agentic" this and that. I kept finding myself unable to clearly explain what was actually happening in these systems, or where to look when something broke.
So I took a different approach: reverse-indexed from typical "intelligent" agent behaviors and asked "what code patterns actually implement this?"
The result is 10 elements that map behaviors like "remembers what I said" or "works while I sleep" to concrete code – conversation arrays, cron jobs, retrieval pipelines, etc.
It's a conceptual map, not a how-to guide. Hopefully useful for anyone trying to make sense of this space.
Elements 8-10 are still drafts. Feedback welcome.