Steven G. Harms on Nostr: It just dawned on me why I'm so scared of Agentic AI. You need to describe something ...

It just dawned on me why I'm so scared of Agentic AI. You need to describe something like a goal and the steps to that goal.

That means you're going to write a DSL for describing steps at the top level and then you're going to have to write definitions of how to do a step.

And, of course, then there's going to be accounting for the fuzz factor in interpreting an LLM's response.

In sum, we're going to be writing #Cucumber tests that can affect the real world.

Unrelated

#ruby #ai