841: AI Vision, Agents and Business Value — with Andrew Ng

Cost Efficiency

Andrew Ng emphasizes the importance of prioritizing the development of valuable AI applications over cost concerns. He suggests using the best available models to create something impactful before focusing on cost optimization. Ng notes that while costs for AI models have decreased significantly, the primary challenge remains building effective solutions.

Use the best model, build something that works, and only after you succeed at that, and only if it turns out to be expensive, then work on the cost optimization afterward.

--- Andrew Ng

Jon Krohn agrees, highlighting the potential of agentic workflows to enhance efficiency without necessarily increasing costs 1.

   

Multi-Agent Systems

The discussion on multi-agent systems reveals their potential to enhance AI functionality by integrating diverse approaches. Andrew Ng reflects on the evolution of AI, noting the blend of Marvin Minsky's Society of Mind theory with modern deep learning techniques. This integration allows AI models to specialize in various tasks, enhancing their overall capabilities.

The reason large language models can demonstrate such amazing capabilities is... because of the richness of the data we feed it.

--- Andrew Ng

Jon Krohn adds that agentic AI workflows can outperform more advanced models in specific tasks, emphasizing the importance of building effective applications 2.