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

Image Revolution

The image processing revolution is being driven by advancements in vision AI technology. Andrew Ng highlights how large multimodal models are transforming industries by interpreting images more effectively. He explains that using agentic workflows, like those in Landing AI's Vision Agent, allows for more accurate image processing by generating plans expressed in code. This approach lowers the barrier for developers to tackle high-stakes visual AI tasks.

Large multimodal models are at least today, they're kind of okay at interpreting images.

--- Andrew Ng

Jon Krohn notes that these advancements are not limited to traditional sectors like manufacturing and healthcare, but also extend to areas such as robotic automation and security, showcasing the broad potential of visual AI 1 2.

   

AI Strategies

Implementing successful vision AI projects requires strategic planning and the use of multiple tools. Andrew Ng emphasizes the importance of agentic workflows, which involve writing code to express plans and test them, thereby simplifying complex visual AI tasks. He advises companies to focus on building valuable applications using these workflows rather than investing in more powerful models.

For most businesses, there's so many opportunities to build applications.

--- Andrew Ng

Ng points out that the cost of using generative AI models is decreasing rapidly, making it more accessible for businesses to leverage these technologies for significant value creation 3.