In the past few years, agentic AI systems like AutoGPT, BabyAGI and others have demonstrated, through prompting, that large language models (LLMs) can accomplish a wide range of tasks, including planning, reasoning and executing multistep processes with little human involvement. They have shown that LLMs can do far more than simply answer individual questions and are capable of taking on more comprehensive autonomous problem-solving.
The appeal has been both rapid and intense. Agentic systems promised lower levels of human oversight, adaptive…
