Google I/O 2024 is one of the biggest software events organized not only for key developers but also for the people using Google workspace around the globe. This event took place on Tuesday, 14th May 2024 at Shoreline Amphitheatre in Mountain View, California at 10am PST, 1pm ET and 6pm BST. Google and Alphabet CEO Sundar Pichai kicked off the event with the company’s vision in the field of AI and its future perspectives. The tech giant also made various announcements launching various AI models that can enhance user experience by simplifying work using AI with the help of tools like Veo, Imagen 3, AI Overviews, Gen AI, Ask Photo and much more. Along with this the integration of the updated version of Gemini into Google Workspace will make it easier for people to complete tasks using AI. Read the story to know about the key features and announcements at Google I/O 2024 in detail.
Google I/O 2024 Highlights
- Google I/O took place on 14th May, 2024 at Shoreline Amphitheatre in Mountain View, California at 10am PST, 1pm ET and 6pm BST.
- The event was organized for less physical attendance but more virtual attendance.
- The event focused more on AI related announcements and less on hardware related announcements.
- This event was one of the biggest events unveiling some of the biggest announcements in the field of AI.
- The tech giant Google has announced the integration of Gemini model into its workspace, therefore working on its workspace with AI powered features like smart reply for emails, searching in google photos and much more, enhances user experience.
- Google and Alphabet CEO Sundar Pichai has announced that cloud customers will get 6th generation of TPUs Trillium in late 2024 which offers 4.7x performance enhancements.
- As per the announcements, Google’s AI has unveiled top features and is competing with Open AI’s features directly.
- Google CEO Sundar Pichai said, “Making AI helpful for everyone” is Google’s “Ultimate” Goal.
- While Google’s ultimate goal is to make AI helpful for all, an important question often-overlooked is: What is the ultimate goal of human life? Can artificial intelligence replace spiritual intelligence? This article not only examines the latest developments in AI technology, but also revives the forgotten facts about human existence.
- Let us explore the details of the AI generated models unveiled by Google at Google I/o 2024.
Ask Photos: Search Within Google Photos
- Ask Photos is an AI powered tool integrated with Gemini which will be used to search within photos with the help of keywords or prompts.
- Google Photos was the first AI powered application by Google which lets you search images by people, pets or places as filters.
- Our Smartphone’s galleries grow with each passing day with photos and videos and sometimes it becomes difficult to search for images or videos which were taken a long time ago. But now with this AI powered tool Ask Photos it is easier for a user to recall a specific memory with the help of a keyword.
- For instance if you type ‘Show me the photo from each monument and museum visited by me’, Google Photos will show you the photos including your visit to the museum or any monument. It saves your time for scrolling.
- It’s going to be rolled out in a couple of months for users to experience the power of AI integration with Google Photos.
Musicfx: Google Made Music Easy to Learn
In collaboration with artists like Jacob Collier, Google has transformed the music industry with the help of AI. This feature powered by AI can help new artists as well as experts in building up their career as it generates music by combining genres, instruments and musical stories by taking text as its prompt. It is a good platform for the beginners to begin their career but one thing to keep in mind is that musicfx also has an embedded digital watermark called SynthID which can easily be identified as an AI generated music.
Gemini 1.5 Flash: Lightning Fast AI Model
- Google Deepmind CEO Demis Hessabis has introduced Gemini 1.5 Flash along with updating Gemini 1.5 pro at Google I/o 2024.
- Speed and Efficiency Optimization are the key features of Gemini 1.5 Flash. It works on the principles of Gemini 1.5 pro.
- It is also capable of understanding multimodal inputs and Google’s breakthrough long context window.
- It can help a user in summarization, video and image captioning, data extraction from large datasets and tables, and can work as a chat application.
- It is capable of doing tasks like Gemini 1.5 pro just because it is being trained by it only. Training is a process in which the most essential skills of the model are being transferred into the smaller, efficient model by the process of ‘distillation’.
- Google has announced to expand the tokens for Gemini 1.5 pro to 2 million for its developers.
Project Astra: AI Assistant
- Project Astra is a universal AI Agent which can be used and can be very helpful in everyday life. Google wants everyone to benefit with AI and hence they have introduced Project Astra.
- It can process multimodal information, understand your context and can give you a suitable reply as in conversation. It is capable of hearing and remembering things and responding to it when the thing arrives as if it is a real time assistant.
- It is capable of detecting almost anything in the camera and fills or suggests whatever is being asked for. Also, it is capable of detecting places around us while looking at it through the camera.
- Demos of it are being shown at Google I/o 2024 and it can be the future of AI Assistant.
- The important thing about Project Astra is that you won’t feel that you are talking to a non-human entity. It looks completely natural.
- With voice modifications, added emotions it all feels natural with Project Astra.
Imagen 3: Text to Image Generation AI Model
- Imagen 3 is an AI driven text to image model that can generate accurate and detailed images.
- Other Text to image models need specified prompts and prompt engineering to perform image generation tasks but Imagen 3 has a better prompt understanding to generate astonishing results according to Google.
- The required camera angles and lenses can be added into the prompt to generate the desired result.
- Images generated from Imagen 3 are higher quality images with richer lightning and better text rendering.
- One can sign up to ImageFX to access Imagen 3 as soon as it is available for the users.
Veo: Text to Video Generation AI Model
- Google unfolds Veo at Google I/o 2024, one of the most promising video generation tools ever produced. With the capabilities of producing 1080p videos, this tool can generate videos of more than a minute in cinematic, visual styles with a huge wide range (timelapses, drone shots, aerial shots and much more).
- This AI powered tool by Google can make video production accessible to everyone. It will create possibilities for growth for a beginner as well as for an expert.
- Veo has capabilities of editing an input video with a given prompt and creating the desired result. Further, it can make an image move by creating a video from an image. Along with this it has capabilities of generating or editing videos without disturbing the key frames.
- Veo has capabilities of a better prompt understanding and prompt to output ratio is accurate. Demos shows that it can be the future of video generation through AI.
- It has an overall better vision and understanding as seen from the demos. Users can access Veo through VideoFX. They need to sign up to VideoFX and join the waitlist before the feature rolls out.
Generative AI & AI Overviews: Enhance Google Search
- Being the leading search engine for years, Google has reimagined the power of a search engine with generative AI.
- Now with Generative AI, search engines can do more than search, it can research and plan things for users as per requirements.
- This is made possible with its integration with Gemini and its approach of planning, multi step reasoning and modality.
■ Also Read: OpenAI’s ChatGPT-4o Breakthrough: A Turning Point in AI Evolution
- AI overview is another search optimization feature powered by Google AI Gemini. Being powered by Gemini, good results are expected.
- AI Overviews can bring out the desired result without even typing the complete query in the search box.
- With AI Overview and Generative AI, a user doesn’t need to split its query into multiple searches, AI can do this in a single query.
- For instance, if you need to ask “find a reasonable two bedroom apartment in Rohini, Delhi, India and show the details of the price it offers along with the time to reach the nearest metro station”
- Users can adjust their AI Overview according to their requirements.
- The Side Panel in Google Workspace is also being integrated with Gemini AI by Google. This ensures the user to access all saved information throughout the workspace.
- Liz Reid (Head of Search at Google) said “Whatever’s on your mind, whatever you need to get done, just ask, and Google will do the googling for you.”
Chip: AI Powered Teammate
- Google unfolded Chip, an AI Gemini powered teammate integrated in Google Chats on a multi user platform.
- Earlier, chats were possible with a single chatbot or an avatar but now with Chip, a chatbot can have a unique identity amidst a multi-user environment.
- Additionally, Chip can respond to any task being allotted to him and can respond to messages and emails being sent to his address.
- Moreover, Chip can build a data storage based upon the information being shared by the team. And can share that information back whenever required.
- Every other team member can see the messages shared by Chip and can see Chip as a member of the team as well.
- This is the first time Google has developed this sort of chatbot that can respond in multi-user workspace using AI.
Gemini Nano & Gemini Gems
- Google has introduced Gemini Nano- an AI that doesn’t require a data network to work.
- Gemini Nano is a lightweight huge language model that lets people create social media posts, create product reviews and write promotional emails along with other blurbs.
- Google is planning to run Gemini Nano locally on Chrome. Microsoft also introduced its AI assistant Copilot but it is incapable of running locally.
- Google is also planning to run it on Devtools which developers use to debug and tune their apps. It will also provide suggestions on how to fix coding issues as well. Along with this Google has planned to introduce it in Pixel series smartphones.
- Google has also introduced Google Gems to Gemini advanced subscribers which lets you create a customized chatbot and allocate tasks to it accordingly.
- Google Gemini calls it a Gem. For example, Gemini can create a coding partner for you to advise on coding at regular intervals. The role and the task of the Gem depends totally on the user.
Introducing PaliGemma and Gemma 2.0
- Google has a family of lightweight models called the Gemma family. These models are used by the developers to integrate AI into different apps and devices. This doesn’t require much input and less power consumption.
- Earlier Google has launched two open source models in it having 2 billion and 7 billion parameters.
- At this Google I/o developer conference, it has launched Gemma 2 series with a lightweight model with 27 billion parameters.
- This lightweight model is expected to offer more accurate results to the developers.
- Vice President of Google Labs, Josh Woodward said “So this 27B size, we’ve picked it intentionally,”
- PaliGemma is a relatively small and lightweight state of the art model developed by Google capable of combining image and text processing for generating text outputs.
- It has an understanding of the context and the nuance and can generate human-like language.
- It has SigLIP-So400m as the image encoder, Gemma-2B as the text decoder.
- It is capable of doing image caption, visual question answering, entity detection and referring expression segmentation.
Unveiling SynthID & LearnLM
- Today people are moving towards the generation of images from AI. AI generated images can also be used to spread misinformation as well. Therefore to tackle this SynthID has been launched in collaboration with Google Cloud.
- It is a tool used for watermarking AI generated images. It embeds a digital watermark directly into the pixels of the image.
- This watermark is not visible to a human eye but can be detected for detection of an AI generated image.
- Photos generated by Imagen use this to watermark images.
- LearnLM is a personalized AI Tutor which is designed to help students in their educational work.
- Google is incorporating LearnLM into its search, YouTube and other areas to help students resolve educational issues.
- Google has partnered with Columbia Teachers College, Arizona State University and Khan Academy for testing of this educational AI model.
Conclusion
After multiple announcements by Google, both Open AI and Google stand edge to edge with each other in the race to become tech AI giant. Both announcements came one after another and have taken the world of AI to next level with the level of outputs produced by them.
Google will be releasing these features very soon and it will be interesting to see in the coming months which tech giant will be rewiring the internet with its features.
Spiritual Intelligence overrules Artificial Intelligence
The lightning pace of advancement in science and technology has not only streamlined our tasks, but also fostered an irresistible habit of searching for solutions to all our problems on Google.
While Google can offer solutions to everyday issues, it cannot replace the need for an immaculate spiritual guide, who has solutions to all problems in life.
Science purists place science above God, however, we cannot ignore the fact that science fails to explain why a dead body cannot be revived, or who exactly breathes life into a mother’s womb?
A Complete Saint possesses the power to transform the entire world and lead everyone to righteousness and complete salvation. Majority of our problems arise from forgetting God. A True Saint bestows us with the authentic worship of the one Supreme God, absolving our sins and leading us away from all sorrows of life.
Sant Rampal Ji Maharaj is the solitary True Saint in this world who provides worship in alignment with the original directives of the sacred scriptures spanning diverse religions. He enlightens us with the correct interpretations of those scriptures and revealing the true essence of human life.
AI assistants can assist us in day-to-day tasks, but they cannot help you fulfill the primary goal of human life. Discover the true meaning of human life, along with path-breaking spiritual knowledge as elucidated by Sant Rampal Ji Maharaj on:
- Website: www.jagatgururampalji.org
- YouTube: Sant Rampal Ji Maharaj
- Facebook: Spiritual Leader Saint Rampal Ji Maharaj
- Twitter: @SaintRampalJiM
FAQs
Answer: The annual developer conference Google I/O 2024 was held on 14th May 2024.
Answer: Google I/O stands for its new generation of innovations implying they can turn any kind of input into any kind of output.
Answer: The event lasted for two days.
Answer: It can be watched on the official YouTube channel of Google.
