Gemini AI will be able to partially control smart home devices in Google Home

This month, Google announced the integration of its Gemini AI with the Google Home app, making it much easier to control smart devices for Android users. Thanks to this integration, users can give commands to Gemini in natural language and the AI ​​will interact directly with the connected devices. For example, you can say “Prepare the bedroom for bed” and Gemini will start the robot vacuum or “Prepare the dining room for dinner” to automatically dim the lights and turn on the music.

Previously, such commands were performed through Google Assistant, but now Gemini is integrated directly with Google Home, which speeds up the processing of commands. At the moment, the control of lighting, climate control and multimedia devices is supported, but the function is not yet available for working with cameras and locks.

To take advantage of the new feature, users need to join the Google Home Public Preview, which provides early access to the features.

Last month, the free Gemini tier was updated to Flash version 1.5, which now processes answers 50% faster. This is made possible by the significant latency improvements that Google has implemented recently. One of the key updates was the increase of the context window to 32 thousand tokens compared to the previous 8 thousand. This means that the Gemini system can now process and store much more textual information when responding to user requests.

Gemini 1.5 Flash was announced to developers in May when Gemini Advanced moved to version 1.5 Pro. In parallel, the Google Tasks extension is starting to expand beyond the Pixel 9 series. It can already be found on devices such as the Pixel 8, although availability to all users is still limited. Among the new features is the ability to take photos of checklists and add them to Google Tasks.

Google also introduced a new @OpenStax extension for Gemini that provides answers based on Rice University’s OpenStax educational resources. In addition, Gemini can now create interactive practice tests on various topics, complementing the platform’s educational features.