AMD CEO Lisa Su on the AI revolution

Chip Design

AMD is leveraging AI to enhance its chip design processes, focusing on both engineering and non-engineering functions. Lisa Su explains that AI aids in chip design, test generation, and manufacturing diagnostics, although human oversight remains crucial 1. She emphasizes the importance of experimentation and rapid learning cycles to integrate AI effectively into AMD's development processes. Nilay Patel notes that while AI is increasingly used within AMD, it won't replace human engineers anytime soon 2.

We have lots of engineers. I think copilot functions in particular are actually fairly easy to adopt. Pure generative AI. We need to check and make sure that it works.

--- Lisa Su

The company is committed to exploring AI's potential while ensuring quality and reliability in its chip design.

   

Software Developments

AMD is advancing AI software by focusing on hardware agnosticism, allowing seamless integration across different platforms. Lisa Su highlights AMD's qualification on PyTorch 2.0, enabling code to run on AMD hardware out of the box, promoting a competitive environment where the best chip wins 3. She advocates for an open software approach, contrasting with proprietary systems like Nvidia's CUDA, to provide developers with more choices and flexibility 4.

The benefit of an open approach is that there's no one company that has all of the ideas.

--- Lisa Su

This strategy aims to foster innovation and collaboration within the AI ecosystem, ensuring AMD remains at the forefront of AI software development.

   

Cost and Value

The economic implications of AI are significant, with high costs currently associated with cutting-edge GPU development. Lisa Su acknowledges the expenses but emphasizes the productivity gains AI offers, particularly in speeding up chip design and testing processes 5. She envisions a future where AI is ubiquitous, operating not only in the cloud but also at the edge, such as in PCs and phones, making it more affordable and private.

Time is money in our world.

--- Lisa Su

This vision of "AI everywhere" aims to democratize access to AI technologies, ultimately reducing costs for consumers.