HIGHLIGHTS
- It has been reported that Google is developing “Pixie,” a new AI helper designed just for Pixel devices.
- According to rumors, this on-device AI would run on Gemini and be capable of tasks that Google Assistant is not.
- In 2024, this might be included in the Google Pixel 9 series.
According to reports, Google is developing “Pixie,” a brand-new AI helper, just for Pixel devices. Pixie, in contrast to the existing Google Assistant, is rumored to be built on Google’s Gemini Nano large language model, which is included in the Pixel 8 Pro. One noteworthy benefit is Pixie’s on-device capabilities, which enable her to complete intricate tasks even in the absence of an internet connection.
Pixie’s Unique Features
According to a recent report by The Information, Pixie is expected to leverage the on-device processing capability of Google’s Gemini Nano, hence offering a more efficient and feature-rich experience in comparison to Google Assistant. Pixie stands out in terms of offline capability because users won’t be dependent on an internet connection to complete chores.
Pixel Model Availability
Although it’s still unknown which Pixel models will get Pixie, there are rumors that it could make its debut with the Pixel 9 series, which is expected to launch early next year. Users who are interested in Pixie may have to wait as Google typically releases new flagships in October.
Exclusivity to Pro Models?
Pixie may only be available on the Pixel 9 Pro, according to rumors, given Google’s propensity to save sophisticated features for its Pro versions. Examples of these capabilities are Summarise in the Recorder app and Smart Reply for Gboard on the Pixel 8 Pro. This exclusivity tactic fits nicely with Google’s previous methods of adding high-end functionality to their more expensive devices.
Gemini AI from Google is Currently Available to Developers
In addition, Google just introduced Gemini Pro, a potent language model intended for developers and businesses. Its goal is to assist students in developing sophisticated answers for a range of problems. The first version features a 32K text context window and can be accessed via the Gemini API. There will be an expansion of this window in later editions.
