'%20x='0'%20y='0'%20height='100%25'%20width='100%25'%20%0A%20%20%20%20%20%20%20%20%20%20xlink%3Ahref='data:image/jpg;base64,/9j/2wBDAAYEBQYFBAYGBQYHBwYIChAKCgkJChQODwwQFxQYGBcUFhYaHSUfGhsjHBYWICwgIyYnKSopGR8tMC0oMCUoKSj/2wBDAQcHBwoIChMKChMoGhYaKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCj/wgARCAAGAAoDASIAAhEBAxEB/8QAFgABAQEAAAAAAAAAAAAAAAAAAAUH/8QAFQEBAQAAAAAAAAAAAAAAAAAAAAL/2gAMAwEAAhADEAAAAN3oif/EAB4QAAICAgIDAAAAAAAAAAAAAAERAgMEBQATIXPh/9oACAEBAAE/ANnXuMoE63KoxSJ1ygZAzQD7AQkX45LKkyqgvZ85/8QAFxEAAwEAAAAAAAAAAAAAAAAAAAERAv/aAAgBAgEBPwBKZh//xAAWEQEBAQAAAAAAAAAAAAAAAAABACH/2gAIAQMBAT8AHb//2Q=='%3E%3C/image%3E%3C/svg%3E)
Google I/O 2024 just wrapped up, and there were some interesting announcements made at the event. The keynote, hosted by CEO Sundar Pichai, unveiled several advancements in Gemini and announced new models like Project Astra, Veo, and more. We saw several improvements in Gemini 1.5 Pro, with its integration into Photos, Workspace, and multiple other Google software, giving full support to the ecosystem.
In this article, we'll uncover all the new announcements made in the Google I/O event.
Gemini 1.5 Pro, Project Astra, Veo, and more were unveiled in the new Google I/O event
1) AI Overviews
AI Overviews uses generative AI to answer search-related queries in Google Search.
Trending
Generative AI is not just limited to text but can also create images, audio, and video. In short, it offers quick answers to complex questions without needing to visit multiple websites. It provides a short general summary of your search query, thus saving time and making you more efficient.
Moreover, Generative AI provides more organic traffic to informative sites and offers follow-up questions to explore topics further. Google promises AI Overview access to over a billion users by the end of this year.
2) Google Photos integration
Gemini will be integrating with Photos to help you search more efficiently. Instead of typing in your search query and then looking through hundreds of similar images, Gemini uses repetitive images as information and answers queries based on the data.
For example, if you forgot the number of your license plate, you could simply ask Gemini, and it'd browse through your gallery and provide results in seconds. This feature is called "Ask Photos with Gemini."
3) Gemini 1.5 Pro
Google introduced the new Gemini 1.5 Pro in the Google I/O event, and it is said to feature a two million token context window. This means it can handle more tasks and process more information. It would allow you to process large codebases, documents, and even audio at a large scale. You now have more insightful answers and informative responses compared to the publicly available version.
However, the public version has a context window of one million tokens, which is still fast. The 1.5 Pro will only be available to developers in a private preview. You can access it through Gemini Advanced.
4) Gemini 1.5 Pro in Workspace
Workspace is another software in relation to Gemini integration, and users will largely benefit from it. Using it, you can summarize emails, get meeting highlights from Google Meet, prepare responses for emails, and much more.
You get access to Gemini in the sidebar of all Workspace apps like Gmail, Drive, Docs, Sheets, and Slides. Gemini 1.5 Pro is available on Workspace Labs and can be tested out.
5) Project Astra
Project Astra uses videos to answer search queries. You simply have to turn on your phone's camera to ask a question, and you will receive answers immediately. This can be anything from a complex math problem, line of code, or even data structures.
The model is also capable of remembering objects from the video and can be used to remind you about something you forgot.
6) Music AI Sandbox
The next feature unveiled at the Google I/O event was the Music AI Sandbox, a model that can create new instrumental segments from scratch. It helps mix different music styles with just a prompt, allowing artists to expand their creative capabilities.
7) Veo
Veo is a new generative video model introduced by Google at its Google I/O event. With its help, you can create and edit high-quality videos by just typing a text, posting an image, or via a video prompt. You can get over 60 seconds of video with different visual styles. From animated shots to photorealism, Veo can make it all, and the only limit is your creativity.
You can join the waitlist in Google Labs to try out the model in the VideoFX tool.
8) Imagen-3
Imagen-3 is a model designed to create images via text prompts.
Similar to Veo, which creates videos off of text, Imagen works with images. It is the company's most advanced model for images and is capable of producing photorealistic photos with the best clarity. It will be available in AI Labs soon but will be focused on developers at first.
9) Trillium
Trillium is the sixth-gen Tensor Processing Unit or TPU. It is a custom-designed chip for machine learning applications.
Trillium offers a significant boost over the last generation, capable of accelerating AI tasks and applications. It will be available to Google Cloud users towards the end of the year.
10) Gemini 1.5 Flash
Announced at the I/O event, Gemini 1.5 Flash features a more speed- and efficiency-optimized interface. It is ideal for high-volume, frequent tasks and has faster response times.
Compared to the 1.5 Pro, which takes longer to generate responses, the Flash is a quicker alternative and requires less computing to run. The Flash has a context window of one million tokens and is publicly available.
11) Gemini Nano
Gemini Nano is another feature introduced at the Google I/O event and will be featured on the Pixel phones later this year.
Now, these phones are not only restricted to text input but can also understand pictures, videos, sounds, and spoken language. This broadens the reach of information attainable by your device and can further help in your search queries.
11) Learn LM
LearnLM uses Gemini to help with educational research. It is aimed at students and provides step-by-step guidance to help you have a better understanding of challenging concepts.
LearnLM will be launched in Gemini soon. Learning Coach will provide insightful answers rather than straight-to-the-point responses, thus making the whole process a lot more engaging and immersive.
These were the major announcements made at the Google I/O event.
Follow Sportskeeda Tech for more info on the topic.
