Hardware limitations: Gemini Nano and the Pixel 8
Google confirmed this today on the Android Developers channel on YouTube, in an episode of The Android Show. In this episode, representatives from the company shared the latest from MWC 2024, Android 15, and more information on Gemini Nano. When the Gemini Nano subject was broached, the company confirmed that Gemini Nano wouldn’t be coming to the Pixel 8, citing “hardware limitations” as the reason. The company did hint at a wider expansion to “high-end devices in the near future,” giving users some hope that at some point, their premium smartphones would benefit from these AI features. However, it appears that non-Pro Pixels, such as the Pixel 8, do not meet the necessary specs to run the powerful AI model on-device.
One of the biggest differences between the regular Pixel 8 and the Pixel 8 Pro is the amount of RAM it has. The Pixel 8 has 8GB, while the Pro boasts 12GB, likely making all the difference for running Gemini Nano. Interestingly, this news comes on the heels of an update to the MediaTek Dimensity 8300 and 9300 chipsets, which are now optimized for Gemini Nano, and could possibly provide mid-range non-Google devices a chance to have these AI features integrated as well.
Currently, AI features found on the Pixel 8 Pro that use this model are the Recorder app’s summarize tool, Smart Reply in Gboard, Video Boost, Night Sight in Timelapse mode, improved portrait light in Google Photos, and the improved Photo Unblur. Having these features available only on Pro models would help set them apart from the regular Pixel line and therefore provide a reason for users to choose the larger model when purchasing.
It will be interesting to see what Google’s next move is going to be as far as rolling out exclusive AI features on future devices. Will the Pixel 9 suffer the same fate as the Pixel 8, and will users have enough of an incentive to upgrade to the Pixel 9 Pro? We will have to wait until the next Made by Google event to find out.
#Google #Gemini #Nano #coming #regular #Pixel