Yeah, my job is now automating tasks using AI and code instead of automating tasks with just code. The code part has gone down, and the abstraction has moved up a layer, but I have more work than ever.
The makes you an an observers orchestrator in a rudimentary agentic system. What happens when that layer of abstraction for scheduling, managing and instructing of agents is better suited for an LLM or group of LLMs? I think a lot of people are going to get screwed.
