AutoGBT can't ask follow up questions, apart from asking for permission to perform its next step. With an autonomous agent, there's no chance to intervene until all of the very long list of tasks is completed. Avram also tried baby AGI, and he basically pointed out something similar to what many have reported with this idea of endless looping.
For the last 3 weeks, AutoGPT has massively captured the attention of the AI community. But how useful is it really? Some are starting to ask whether it really lives up to the hype. Watch the original video: https://www.youtube.com/@TheAIBreakdown