Google launches new AI at Google I/O with innovative new features

Google announced the new update at Google I/O 2024, Google's annual technology launch event, at midnight in Thailand. Google's announcements of AI updates like Gemini Pro 1.5, new Android updates, Gemma2 updates, and many more have created quite a stir. and ripples within the AI ​​circle, which is only a short time away from the launch of GPT- 4o. The Techcrunch website has summarized the update announcement with the following details:

Firebase Genkit

This update to the application development platform makes it easier for developers to develop AI-powered applications in JavaScript/TypeScript, with support coming soon. The open-source Apache 2.0 license enables AI developers to quickly build AI into new or existing applications.

LearnLM

Google launched LearnLM at the event, a next-generation AI model optimized for learning. A collaboration between Google's DeepMind AI research division and Google Research, the model is designed to teach students a variety of topics in a conversational format. Currently, Google has introduced LearnLM to be tested on various platforms and is currently piloting it in Google Classroom.

Quiz master

A prototype update for YouTube is the creation of AI-generated quizzes. This tool allows users to ask users comparative questions when watching educational videos. Video viewers can ask questions and receive helpful explanations. Or take a quiz about the content within the video.

Gemma2's update

As a result of developer requests for a larger Gemma model, Google will be adding a new model with 27 billion parameters to Gemma2. This version of Google's Gemma model will launch in June. Optimized by Nvidia to run efficiently on next-generation GPUs on a single TPU host and AI zenith, Google said.

Google Play

A new search feature has been launched for the application. It provides user acquisition, Play Points updates, and other improvements for developers such as Google Play SDK Console, Play Integrity API, and many more. More interesting feature updates for developers called The Engage SDK will guide app creators on how to display their content to users to create a full-screen experience. and adjusted to suit each user as well

Call center number detection

This is another update that will notify users of fraudulent calls during calls. This feature will be rolled out to Android in the future using Gemini Nano, which is a lightweight version of Google's original AI plan so it will be available across all devices. The system will listen for suspicious conversation patterns in real-time. to identify fraudsters

Ask Photo

Another new feature from Google Photo that is powered by the AI ​​engine Gemini, this extension will be available later this summer. This feature will allow people to search their Google Photo collection using natural language, leveraging an understanding of image content and other information to identify locations, people, or objects in images. with the LLM model

Gemini's big update

You can use Gemini in Gmail to search, summarize, and draft emails.

-Gemini 1.5 Pro is an update from the previous version that adds features such as document analysis, code base, video, and audio recordings that last longer than the previous version.

-Gemini Nano, the smallest model of Gemini available directly in Chrome, will launch in Chrome 126, which Google says will help developers use the models on their devices to power AI features.

-Gemini in Android, which replaces Google Assistant, will take advantage of the ability to integrate with mobile systems to increase usability.

-Gemini in Google Maps, which can be made available to developers starting from the Places API. Developers can display AI summaries of places and regions. Summary of in-depth information etc.

Introducing a new processing chip

Google has announced a new AI chip with six Tensor Processing Units (TPUs) called Trillium, set to launch later this year.

Upgrade Imagen 3

Google today announced Imagen 3, the latest version of its visualization model. Demis Hassabis, CEO of DeepMind, Google's AI research arm, said Imagen 3 more accurately understands text prompts to be translated into images. Compared to its predecessor Imagen 2, it is more "creative and detailed" than its predecessor.

Project IDX

The AI-centric developer space in the browser is now in beta. With this upgrade, Google Maps will be integrated into the IDE to provide additional capabilities, as well as integration with Chrome Dev Tools and Lighthouse to help debug applications. Google plans to Make the app compatible with Cloud Run, one of Google Cloud's frameworks.

VEO

Google has launched VEO, an AI model that can create videos with a resolution of up to 1080p, approximately 1 minute long. It can create videos from both text and image styles. and different movie genres It also creates a landscape image. and Time Lapse by Google released a preview at the end of April.

Circle to Search

The AI-powered circle search feature, which allows Android users to instantly display search results using gestures such as circles, will work very naturally with Google Search from anywhere on their phone. You can do things like circle, highlight, draw, or tap on your device.

Pixel products

The Pixel 8a was the first image released at Google I/O after it was announced last week. It will be sold at a starting price of $499 and the Pixel Tablet is already available for purchase.