This year’s COMPUTEX Forum was held on Wednesday with the theme of “Let’s Talk Generative AI”. The morning session brought together four international AI experts to speak to how generative AI is transforming businesses and industries everywhere.
The forum was kicked off by Marc Hamilton, Vice President of Solutions Architecture and Engineering at NVIDIA, with a presentation on the theme of “Infra Build Train Go” detailing how his team are helping clients to quickly utilize and integrate generative AI.
NVIDIA have been building AI factories for eight years already, and Hamilton’s team have overseen more than 250,000 installs of the H100 GPU. Hamilton’s three key metrics here were time to first train, GPU availability, and time to train. Once the factory is built, the company’s NIMs provide developers with ready-to-use AI microservices that allow them to hit the ground running.
Next was John Solomon, Google’s Vice President and General Manager of Chrome OS. Solomon gave a talk that positioned Google’s Gemini model at the forefront of AI development and rollout by emphasizing how its native multimodality and total scalability from the cloud to mobile devices is putting it ahead of the competition.
Gemini is already transforming Google’s user offering in multiple arenas: throughout the company’s popular Workspace suite of tools and applications; integrated natively with Android OS; and in the new Chromebook Plus series of AI laptops. Starting at $349, these laptops are making AI accessible to consumers at competitive price points.
Thomas Anderson, Vice President of AI and Machine Learning at Synopsys, took the stage to explain how his company is building generative AI into its chip design tools, which are used in turn by partners including NVIDIA, AMD, and Qualcomm to design their latest AI chips.
Unlike software design, chip design often requires tens to hundreds of engineers and a development cycle of up to 24 months or longer, so tools like the company’s Synopsys.ai automation suite are crucial for the industry to keep up its forward momentum by quickly solving technical problems and allowing chipmakers to focus on creative, innovative designs for the future.
The morning session was wrapped up by Xia Zhang, Enterprise Strategist at AWS, who presented on the theme of generative AI and data innovation. Zhang began by pointing out that “data is your differentiator for generative AI”, and outlined how AWS are providing data strategies that are comprehensive, integrated and secure.
Solutions are available across the entirety of the stack, from silicon to software. AWS produces Trainium and Inferentia, its purpose-built machine learning chips, and its Sagemaker and Bedrock platforms are widely used for building and scaling ML models and AI applications. Finally, the Amazon Q AI-powered assistant is helping clients to unlock data-driven productivity and capabilities.
