Google has revealed numerous new features and products at Google I/O, although most have yet to be released. Let's review what we currently know about the developments in progress.
Upcoming Gemini Announcements
Google has scheduled five Gemini announcements for July 15 and July 18. While the exact details are still unclear, several strong contenders have emerged.
Disclaimer
Most of this information is based on reverse engineering of the frontend code and leaks related to visible UI components. Predicting new model announcements is much harder.
Imagen3
Aside from VideoFX and Veo, Imagen3, which is expected to be available to alpha testers in AI Labs, might also become available to Gemini Advanced users eventually. While its release next week is unlikely, considering the quality of Imagen2 and the lack of image generation availability in the EU, it is a highly anticipated release. Rumours suggest that the first invites will be sent to AI Labs Discord members and those with Advanced subscriptions.
GEMs
Custom GPTs for Gemini, called GEMs (previously known as “Bots”), have been in development even before the I/O announcement. Users will be able to view, edit, and copy GEMs, accessible from the GEMs Manager tab. Given their long development time, GEMs might be a significant release but could also be delayed.
Memory or Personalized Responses
This feature has been in the code for a while and is expected to function similarly to the Memory feature on ChatGPT. Users will have a dedicated option in settings to access the personalization section. However, as this section is named “Personalized Responses,” it might have a different twist.
Scheduled Prompts
This relatively new addition appears as a separate section in the side menu and the icon behind a Gemini response. The tooltip suggests this button will allow users to schedule prompts. In the dedicated tab, users will likely see a list of scheduled tasks. This unique feature could enable users to ask Gemini to send them daily news every morning, working well in combination with GEMs.
Voice Recording and Google Photos Integration
The attachment option indicates two new additions:
- Voice Recording: Allows users to record a message and send it as a .wav file. While the recording feature seems to work, it feels far from being close to release.
- Google Photos Integration: Appears nearly finished, allowing users to select photos directly from the Photos app on the web. However, it still does not address the concern of not being able to upload multiple images at once.
Prompt Enhancements
A new hidden button might function as a prompt enhancer based on its appearance and name.
Gemini Chrome Extension
There is a menu option saying “Send links in prompts (Chrome Extension),” which might be a developer tool or an experiment that may never be released. If implemented, it could allow users to send links directly from Gemini’s conversation.
Real-Time Responses Toggle
Future updates might include the ability to disable real-time responses.
Additional Features
There are more Gemini features for Android spotted earlier. Additionally, Google is recruiting beta testers from Betabound for their iOS Google app, indicating that iOS updates may follow later 👀
Google Gemini is the latest and most advanced artificial intelligence model developed by Google, representing a significant leap in AI capabilities. Introduced in December 2023 and widely released in February 2024, Gemini is a multimodal AI system capable of understanding and processing various types of input, including text, code, audio, images, and video. It comes in three main versions: Gemini Ultra (the most powerful), Gemini Pro (for wide-scale applications), and Gemini Nano (optimized for on-device tasks). Gemini has been integrated into various Google products, including the rebranded Bard chatbot (now called Gemini), Google Search, and Pixel smartphones. With its ability to handle complex tasks, generate content, and assist in a wide range of applications, Gemini represents Google's ambitious effort to compete with other leading AI models like OpenAI's GPT-4, potentially reshaping the landscape of AI technology and its applications in everyday life.
Source: @testingcatalog