Another thing I started doing was using n8n for the more de…

Another thing I started doing was using n8n for the more deterministic workflows like fetching data, formatting, sending messages , and then the LLM where it matters. I bet that is where this is heading. I bet people are also using smaller models and bigger models for different tasks. I can't imagine an efficient system isn't being worked up right now when people like me sitting here by myself figure that out in a couple days.