Gemini, the on-device AI model, will be available on Google’s Pixel 8 smartphones. The phone will come with Gemini Nano, a version of the model designed to run locally on personal devices. This follows a successful rollout to the Pro model and the Samsung S24 in January.
The Pixel 8 will feature the proprietary Tensor G3 chip, the same as the Pro model, which is meant to enhance AI performance. Therefore, the overall experience on both gadgets should be similar. The rollout will happen in the next Pixel Feature Drop, but initially only as a developer preview. Google wants to gather feedback and ensure everything runs smoothly on the slightly lower-specced phone.
This is a surprising change for Google, considering the initial statement that the Pixel 8 couldn’t support on-device Gemini due to “hardware limitations.” The main differentiating factor between the Pixel 8 and the Pro model is the RAM allotment, which seems unlikely to hinder on-device AI functionality. It appears that Google has reconsidered its stance on this issue.
Gemini Nano will power enhanced features in two tools that utilize the LLM. The Recorder app will receive a Gemini-powered summarize feature that works locally on the device, eliminating the need to send data to a server. This feature is used to create summaries of recorded conversations. Additionally, Gemini Nano will support Gboard’s Smart Reply toolset, which suggests responses to messages based on conversation context. More messaging apps are expected to add support for this feature in the future.
Despite recent controversies surrounding Gemini, Google continues to expand its capabilities. Apple is reportedly exploring the implementation of the model, and Gemini has been criticized for its image generation feature, which is currently on hold for further refinement.
Source link
