[CuriouslyC]

Google doesn't seem to "get" agentic autonomy. Their models are trained to solve short problems really well, but they get confused over long time horizon tasks and kinda suck at tool calling to boot.