Google I/O 2024 was one of the most important events of the year, leaving us on the table with Google’s artificial intelligence offering for the near future. It came just days after the big announcement of GPT-4o and brought us news for Android, WearOS, Android Automotive, and any devices that can run Gemini, Google’s family of artificial intelligence models.
At Xataka, we’ve rounded up all the news presented at Google I/O 2024, from improvements across the model family to all the news coming to Android when the stable version of Android 15 is released.
Google told us about all the improvements that the Gemini family will receive in 2024. First up, we have the Gemini Flash, its lightest model. In the Gemini family, it is the fastest, lowest latency solution running locally on devices (and will be an ideal alternative to Gemini Nano in the future). Those applications that need fast response and not so much a denser model can use Flash.
In addition to introducing the newest member of the family, Google told us about the improvements coming to Gemini 1.5 Pro. The first of them is computation speed. 1.5 Pro will be able to analyze large documents: files totaling up to 1,500 pages, summaries of up to 100 emails at once, an hour of video, or codebases of more than 30,000 lines.
It will also become smarter at understanding images and be able to plan trips in detail in real time based on information from our emails, flights, calendar and more.
He naming With Google this is not an easy task to understand, but let’s go there. When we talk about Google’s family of artificial intelligence models, we talk about Gemini. And Gemini now consists of:
As if that wasn’t enough, Google told us about “Project Astra.” And what is this? A project through which they want to “develop universal AI agents that can be useful in our daily lives.” This is the materialization of Google’s plans to create the perfect assistant that remembers what it sees and hears with low latency and different tones of voice.
This Google “She” will eventually be integrated into Gemini (as will the voice assistant they are testing to replace Google Assistant, which is also called Gemini), so it will eventually become Gemini like everything else. In fact, one of the most interesting integrations we saw was Project Astra running in glasses: a vision of what’s to come.
During I/O, Google also announced I See and Image 3, developments of its text-to-image and text-to-video models. They want to compete with Midjourney, the future Sora and company.
Image 3 is an evolution of Image 2 and has improved natural language interpretation as well as the ability to add very specific details from the image. tips very extensive.
Veo is Google’s largest generative video model to date, allowing the creation of Full HD (1080p) clips over a minute in length. You can also create time interval and is capable of applying all types of cinematic techniques.
Google Search was another one of the I/O heroes. As part of a feature called AI Reviews, Google showed us a feature of its search engine for displaying AI-generated results. This feature will be available in the US first, although we can test it through Google Labs if we use a VPN.
As for Android, there have been news related to both the operating system itself and the applications associated with it.
Image | Google
In Hatak | The movie ‘Her’ is becoming a reality as people begin to develop relationships with their AI bots
Jack Teixeira, a 22-year-old former US military and Air National Guard member, has been sentenced…
Fatima Del Reino Iniesta (Alicante)In Spain, less than 3.5% of candidate patients receive monoclonal antibody…
Neinor Homes Board of Directors announces it will vote to approve the distribution dividends of…
It's been a few months since Stellar Blade launched on PS5, and after a successful…
12.11.2024 19:09. Updated 11/12/2024, 7:16 pm.The scandal over the failure that Argentine football suffered last…
Il ya bien sûr les Scenes d'action déjantées, continuous pyrotechnics, cascades and burlesque comedy. More…