Since the launch of OpenAIs GPT-4 API last month to beta testers, developers have been experimenting with making agent-like implementations of the AI model that attempt to carry out multistep tasks with as little human intervention as possible. Two open source projects, Auto-GPT and BabyAGI, have captured significant attention on social media. Auto-GPT chains together thoughts of the GPT-4 language model to autonomously achieve any goal set. The homebrew script can loop, iterate, and spin off new instances of the AI model as needed. Although not yet as autonomous as promised, these open source projects represent early steps towards more complex chaining AI models that could potentially be more capable than a single AI model working alone. Auto-GPT could serve as an AI assistant to perform any digital task by itself, however, it needs appropriate APIs to be feasible. Someone has created a web-based version of Auto-GPT called AgentGPT that functions similarly.
