Raw LLM Responses

It's easy to regress to certain viewpoints such as "AI will take over jr dev" or the converse viewpoint, "AI is useless". It's the type of stuff that'd easily get upvoted, rather than actual thought into how things can be better utilized and *are* being utilized. Focusing on incremental improvements in model space is focusing on the tree rather than the forest... There's been tremendous innovation in the application space. Many orgs are using agents throughout the org as you said, across multiple verticals. They are beyond useful if you're an expert, and can be reasonable even if not - hence why things like code reviews from more senior members are a thing. We've been able to lower a ton of operational efforts through varying agents across the org. This concretely resulted in a lower heacount increase than we would have had otherwise.