Even for an agentic process, you need a human instrumenting it, to chart a course and review. Whether you're coworking with Claude or running agents, you need to get dirty with the model.
I feel like we're discovering that swaths of humanity have no inner dialogue. If you had the opportunity to work with the world's smartest professor, would you know what to do?
Massive output uptick due to agentic AI. Complete flat adoption.