I RECENTLY joined the Same Brain Podcast to discuss the latest AI trends with popular content creators Justine and Jenna Ezarik. We talked about how AI is changing the way people use their devices and get things done. I also demonstrated new products that run on-device generative AI and showed that artistic talent is not required to create strong results using image generation.
Using AI in our daily lives
Like many people, I use AI tools every day to improve productivity and support creativity, both at work and at home. I told Jenna and Justine how I have used chatbots to invent new cocktail recipes, write birthday poems and help resolve family debates on obscure topics.
Many devices now have generative AI tools built into apps people already use, helping users stay “in the flow.” Chromebooks are one example, with Gemini integrated into tools such as Circle to Search, which makes it easier to get information about what is on your screen, and Smart Grouping, which helps manage and organize multiple tabs.
AI-powered creative tools are also becoming more capable. In the video, I put my artistic skills to the test by walking through image generation features on the Samsung Galaxy Tab S11, which uses a MediaTek Dimensity 9400+ chip with a neural processing unit.
Agentic AI is the next step
Agentic AI is more proactive and personalized than generative AI. It is designed to learn user preferences and combine that information with other data to provide tailored assistance.
On the podcast, I shared an example of how agentic AI could simplify everyday tasks. A chatbot could automatically book a ride to the airport, taking into account your preferred rideshare service and how early you typically arrive before flights. While this may seem minor, Justine noted that reducing the mental load of scheduling and coordination can make a meaningful difference.
Processing information on-device
Generative and agentic AI are largely cloud-based because of the scale and performance required by data centers. However, more processing is moving from the cloud to the edge. One reason is the development of more powerful neural process ing units capable of running large language models with billions of parameters on-device. At the same time, smaller language models optimized for specific workloads are becoming more common and effective at the edge.
On-device processing reduces latency, improves user experience and allows AI systems to better personalize responses while protecting user privacy.
This shift is not limited to smartphones, tablets, Chromebooks and PCs. Nvidia’s DGX Spark is an AI personal supercomputer that allows developers to prototype and fine-tune AI models with up to 200 billion parameters locally. It handles these workloads using the GB10 Grace Blackwell Superchip, co-designed by MediaTek and Nvidia, which balances performance and power efficiency.
Adam King is the vice president and general manager of the Personal Devices Business Unit at MediaTek, a Taiwan-based fabless semiconductor company that designs system-on-a-chip (SoC) solutions used in a wide range of consumer electronics.
