Useful in many ways, Gemini becomes much more. During his annual conference dedicated to developers, Google presented their projects for their AI. “Our final vision is to transform the Gemini application into a universal AI assistant,” said Demis Hassabis, CEO of Deepmind, the subsidiary of Google.
In other words, “this is an intelligent AI, which includes the context in which you are and that you can plan and act in your name, in any device,” he said. Gemini, which will be invited to many devices, should, in the long term, make users more productive, particularly when performing daily tasks for them.
A more powerful and more personal Gemini
A vision still distant, which includes, among other things, two projects in the prototype stage. The first is the Astra project, a multimodal AI that has screening and screen exchange capabilities. The skills that have been integrated into the Vocal Gemini Live, which allows her to “see” through the smartphone camera and answer questions about what she “sees.” Uses that are now available for free for all users.
The Astra project will also be integrated into the search engine, which allows Internet users to discuss what they see in real time with AI.
The second is the Marimer project, launched last December. It is a “agent research prototype” that can help the user “to administer up to 10 tasks simultaneously, from investigations to reservations and purchases.” In other words, it is composed of a network of agents, these AIS capable of carrying out actions instead of the user, which can search for information, make reservations or even make purchases at the same time.
Integrated in the Gemini application, the Marimer project will only be available in the United States at this time. In addition, it will be necessary to subscribe to the new formula, Google AI Ultra, launched on Tuesday, May 20 in the United States. For $ 249.99 per month, it allows you to have less limits in terms of use, but also access the most efficient AI models and premium functions. It is mainly aimed at professionals as directors or developers.
Because Google has presented other functions to make Gemini “more personal, more proactive and more powerful.” During its conference, the company presented several tools dedicated to creation. Among them is Image 4, a new image generation model that “combines speed with precision to create magnificent images.” Depending on the company, it offers remarkable clarity in details such as water drops or animal skins.
I see 3, meanwhile, you can generate quality videos with sound for the first time. Either traffic noises, birds that sing in a park or even dialogue, can be used to tell stories of a simple application. The introductory video of Google I/or was also generated with this AI tool.
Finally, Google presented Flow, a realization tool to create clips, scenes or cinematographic stories with VI, Image and Gemini. Something that will be possible to do with natural language.
All these functions will also be available only for Google Ai Ultra subscribers, as well as the use of Gemini in Chrome, Gmail or Docs. However, some of them will be accessible with Google AI Pro, a new name of the Premium formula of Google One AI (21.99 euros per month).
Source: BFM TV
