Google announced a range of noteworthy developments during its I/O 2024 event. The CEO of Alphabet and Google, Sundar Pichai, outlined the company's AI strategy. Notably, Google updated the Gemini model family with some significant modifications. Project Astra, which sees the use of AI agents in the future, and the new Gemini 1.5 Flash, a lightweight model built for speed and efficiency, are two examples of this. Google's continuous efforts to advance AI and offer tools that improve user experience across various applications are reflected in the latest improvements. Google is anticipated to advocate for the ethical development of AI.
Also Read: Google Pixel 8a with Tensor G3 Chipset Launched in India at Rs 52,999; Check Details
Google is improving Gemini Nano, its on-device mobile large language model. Pichai announced onstage that the tool will now be known as the Gemini Nano with Multimodality, allowing it to “turn any input into any output.” This implies that it can extract data from text, images, audio, web or social media videos, and live video captured by your phone's camera, then combine that data to create a summary or respond to any queries you may have. Google showed a video in which someone used a camera to scan all of the books on a shelf and record the titles in a database so that they could be recognised later. The company's more robust cloud-based AI technology, Gemini 1.5 Pro, is now accessible to all developers worldwide.
Google has also announced the newest iteration of the Gemma AI, which promises increased effectiveness and performance. With 27 billion parameters, the new model has been optimised using TPUs and GPUs. The company introduced its first vision language model, PaLI-3, to the Gemma family of models. The new models will be available in June.
Within Google Photos, the manufacturer has included some advanced visual search capabilities. Gemini can now search your photos and provide more detailed results with a new feature called Ask Photos. For example, if you give it your license plate number, it will search all of your photos for your car using context clues. According to Google Photos software engineer Jerem Selier, the feature does not gather data on your photos that can be used to deliver adverts or train other Gemini AI models (apart from those used in Google Photos).
Additionally, Google is incorporating AI into its Workplace office toolkit. Many Google apps, such as Gmail, Google Drive, Docs, Sheets, and Slides, will now have a button in the side panel to toggle Google's Gemini AI. The Gemini helper can assist with resolving queries, drafting emails or documents, and summarising lengthy emails or documents.
Not to give the impression that all of this is limited to office work, Google showed a few parent-friendly capabilities, such as AI chatbots that can help children with their homework or give you a rundown of any PTA meetings you might have missed. Google's Circle to Search, which made its debut earlier this year, is likewise undergoing an update and will soon be utilised to provide kids with academic assistance, such as an explanation of math problems. A Gemini-powered AI Teammate is also included in apps like Docs and Gmail. It will assist you with creating to-do lists, managing project files, tracking down assignments, and improving team communication.
Moreover, the new Gems feature lets you set up automated routines for tasks you want Gemini to perform regularly. It can be configured to perform a variety of digital tasks when given a voice command or a text prompt. In a play on the Gemini name, Google refers to each of those processes as "Gems."
Google has launched the Gemini 1.5 Flash language model, which it claimed to be faster and more effective than the popular Gemini 1.5 Pro. The new AI model has a larger context window of one million tokens and can reason in multiple ways. The Gemini 1.5 Flash model, according to Google, has been fine-tuned to perform well in high-frequency tasks where response speed is critical.
With the release of Project Astra, a brand-new AI agent, Google squared off against OpenAI. In a Google demo, the AI assistant could name the things in the room, recognise and explain a certain section of the code, locate the object precisely by peering out the window, locate the user's glasses, and even come up with imaginative names for dogs. Furthermore, Google demonstrated how to integrate Project Astra with smart glasses or a smartphone, indicating that Google Lens may get a significant Gemini-powered update in the future.
Introducing Veo, its most recent text-to-video generation model, the tech giant has challenged Sora's dominance in the field. Videos with a range of cinematic and visual styles in 1080p resolution can be produced with the new model. Thanks to its comprehension of the terminology used in cinema, such as "timelapse" and "aerial shots of the landscape," Veo will also allow users more control.
