Agentic Workflows
Adam discusses the concept of LLM workflows, emphasizing the shift towards agentic behavior where LLMs can create their own plans and tools. He contrasts this with Marvin, a prompting library designed for easy interaction with LLMs using Python. The conversation highlights the challenges of observability and debugging in agentic workflows, particularly when LLMs take control and how to manage task dependencies effectively.In this clip
From this podcast
Practical AI
Practical workflow orchestration
Related Questions
What did Marc Andreessen say about AI being the control layer?
How will large language models (LLMs) and AI change software engineering and the software development lifecycle (SDLC) as discussed in the episode Making Your Company LLM-native // Francisco Ingham // #266 and the clip Enhancing, Not Replacing?