Your compass for the ever-evolving world of Open Source AI. We explore cutting-edge GitHub repos, demystify complex models, and spark discussions on the future
@username_null00@bendee983 No you dont need human need to review. The human who is steering the agents need to have an understanding what is happening , flow wise , sydtes wise. He is not traditionally reviewing code. Agent with a conpletely fresh context will be doing that.
@username_null00@bendee983 With no human oversight? Its about outsourcing thinking not understanding. And a lot a production grade application are having way worse code than what any ai will be generating.
@ScottCywinski@bendee983 That is not true. Almost all major coding agents have code review plugins like claude code. There is reason why agents.md and claude.md exists. With time AI based best coding practices will be formalized.
@championswimmer AI agents are 3-5 x better when working with files. If you are building agents just try converting you data into files. You will immediately notice performance boost.