Dongming Lei’s Post

The #GoogleIO keynote is always one of my favorite Google moments of the year: it's a sneak peek into our very latest product and technology innovations, many of which of course are enabled by #AI. Here are some of my highlight announcements: ✨ Gemini can understand more information than ever before: We’re bringing Gemini 1.5 Pro to Gemini Advanced subscribers in more than 35 languages, along with a 1 million token context window — the longest of any widely available consumer chatbot in the world. ✨ Gemini in Workspace: We’re integrating Gemini 1.5 Pro into our Workspace apps, for example into the side panel of Gmail, Docs, Drive, Slides and Sheets ✨ Expanding AI Overviews in Search: With a new customized Gemini model — capable of multi-step reasoning, planning and multimodality — combined with best-in-class Search you’ll soon be able to ask complex, multi-step questions, customize search results, and even ask questions with videos ✨ Gemini for Android: We’ll be building AI right into the Android operating system with Circle to Search and the new assistant Gemini on Android ✨ AI supporting the creative process: With Veo, we’re introducing our most capable model for generating high-definition video and with Imagen 3, our highest-quality text-to-image model ✨ We’re expanding our technical watermarking innovations like SynthID in two new modalities — text and video — so AI-generated content is easier to identify I’m also very excited about Ask Photos: Using Gemini’s multimodal capabilities, you’ll be able to simply ask Google Photos to show you the photos you are looking for – for example all the photos of your daughter’s birthday parties over the last 10 years – without all that scrolling! Read more about our announcements at I/O in Sundar’s blog post below 👇🏻 https://lnkd.in/dmtNC6nE #Google #AI #GoogleIO #Gemini

#Wow, #wow, #wow, what an exciting Google I/O, packed with so many product innovations, news and announcements – and of course #AI! Or, as our CEO Sundar Pichai said during his keynote: “Google is fully in our #GeminiEra”. I just finished watching the show and below are my favorites, what did you like the best? Let me know in the comments! ▶️ Gemini 1.5 Flash and Gemini 1.5 Pro are now generally available, coming with our 1 million token context window by default on Google AI Studio and Vertex AI. And we’re expanding it even to a 2 million context window for developers and Google Cloud customers via waitlist. ▶️ That’s the longest token context window of any widely available consumer chatbot in the world – and it means that soon, Gemini 1.5 Pro will be able to handle an hour full of video content or more than 30,000 lines of code. ▶️ We’ve announced the next generation of open models, Gemma 2, and took a look at future of universal AI agents with Project Astra, a responsive agent that can understand and react to the context of conversations. ▶️ We unveiled the 6th generation of our TPUs, called Trillium, which delivers a 4.7x improvement in compute performance per chip over the previous generation, TPU v5e, and which we’ll make available to Cloud customers later this year. ▶️ We’re developing a cutting-edge technique we call AI-assisted red teaming, that draws on Google DeepMind's gaming breakthroughs like AlphaGo and expanding our technical watermarking innovations like SynthID in two new modalities — text and video — so AI-generated content is easier to identify. Check out all our I/O highlight announcements in Sundar’s blog post here 👉🏻 https://lnkd.in/d46_ARs3 #Google #AI #GoogleIO #Gemini

Exciting updates from Google I/O and Google Cloud Next on the latest Gen AI capabilities and models! Gen AI is rapidly evolving, offering solutions akin to a 'Genie' fulfilling wishes. Key highlights include: - Global availability of Google's powerful LLM Gemini 1.5 Pro for developers, alongside the new Gemini 1.5 Flash language model, offering multimodal reasoning at 95% lower cost than GPT4 & Gemini Pro. Gemini Nano: The smallest, most efficient AI model optimized for mobile devices. Gemini Advanced now with a 1 million token context window and can handle 1,500-page PDFs. - PaliGemma, the lightweight vision-language open model optimized for visual Q&A and image captioning. PaliGemma takes both images and text as inputs and can answer questions about images with detail and context. - 'Project Astra,' the new AI Agent, a visual chatbot powered by Gemini, acting as your Gen AI Assistant engaging users conversationally through phone or smart glass cameras and mics. - Introduction of Gemini AI Teammate, an AI companion managing emails, chats, tasks in Google Docs/Sheets, and web searches independently. - Google Photos' Gemini AI-powered feature 'Ask Photos,' allowing users to request specific photos related to events, objects, or people without traditional filter searches. - AI Search capability with Google Search being a pure Gen AI based Search, providing categorized AI-generated search results under ‘unique’ AI-generated headlines. - VEO, the Text to Video Generation Model similar to SORA, generating 1080p resolution videos with a range of cinematic and visual styles. - Imagen 3, a high-quality text-to-image model, and the Music AI Sandbox, a suite of music AI tools. - Precursor to Google I/O was Google Cloud Next in April 2024, which has introduced the No Code Vertex AI Agent Builder. Most of these capabilities mentioned above are based on multiple gen AI Agents. Vertex AI has a no code platform to build LLM Agents within few clicks , can use orchestrator like Langchain. Users can tap the power of Google's foundation models, search expertise, and conversational AI technologies to create enterprise-grade generative AI applications. Stay tuned for more innovative advancements in the Gen AI landscape! [Link to learn more](https://lnkd.in/gxPEdk4c) #GoogleIO #GoogleCloudNext #GenAI #ArtificialIntelligence #TechnologyInnovation #VertexAI #VertexAIAgentBuilder

Summary of Google's Announcements at Google I/O '24 [Part 1/2] Google announced numerous innovations in the field of artificial intelligence during a live event on Tuesday (May 14). Here are the key highlights: ### Real-Time Accessibility Google emphasized the importance of accessibility with a real-time captioning service. **New Model Named "Masquerade"** * This model generates highly accurate and reliable human faces. **"Bard"** * An experimental AI for conversations, showcasing impressive capabilities in dialogue and interaction. **Music FX DJ Tool** * An AI tool for creating music based on user instructions, demonstrated by improvisational musician Marc Rebillet. ### Innovations in Google Workspace **Gemini** * A new generative AI system with significant impact on products like Search, Photos, Workspace, and Android. * Revamped search experience with AI Overviews to ease information retrieval. * Audio Overviews feature for creating a personalized learning experience. **Additional Updates** * A new image model named Imagen 3. * A new generative video model named Veo. * The sixth generation of TPUs, Trillium, offering significant performance improvements. ### AI Tools for Businesses and Individuals **New Features in Gemini** * Summary cards for emails. * New Q&A capabilities. * Plans to expand Gemini to 35 languages, making it more multimodal, adaptive, and intelligent. **Additional Tools** * A new approach to complex problems and extracting information from PDFs. * Improvements in availability and fraud detection tools. * New models for developers like Gemini Nano and Gemini in Android Studio. Hadas Frank #google #genAI #ai

𝐆𝐎𝐎𝐆𝐋𝐄 𝐅𝐈𝐆𝐇𝐓𝐒 𝐁𝐀𝐂𝐊 ! Following the recent launch of GPT-4o by OpenAI, Google has unveiled an array of groundbreaking Generative AI products at their latest Google I/O conference. Here is a glimpse of what is coming: 🌐 𝗚𝗲𝗺𝗶𝗻𝗶 𝟭.𝟱 𝗣𝗿𝗼, 𝗔𝗱𝘃𝗮𝗻𝗰𝗲𝗱, 𝗮𝗻𝗱 𝗙𝗹𝗮𝘀𝗵: 𝘎𝘦𝘮𝘪𝘯𝘪 1.5 𝘗𝘳𝘰: Now available globally to developers, this version supports up to 2M tokens in context window, enabling the model to process extensive content such as entire TV show scripts. In addition, it can summarize entire email conversations and Google Meet recordings in Workspace Labs. 𝘎𝘦𝘮𝘪𝘯𝘪 𝘈𝘥𝘷𝘢𝘯𝘤𝘦𝘥: Available in 35 languages, this model supports a 1M tokens in context window. 𝘎𝘦𝘮𝘪𝘯𝘪 𝘍𝘭𝘢𝘴𝘩: A lightweight version designed for reduced latency applications, also supporting 1M tokens. 🤖 𝗣𝗿𝗼𝗷𝗲𝗰𝘁 𝗔𝘀𝘁𝗿𝗮: Google’s answer to GPT-4o is Project Astra, a universal AI assistant that understands context and responds with human-like emotions. Look out for its integration into the Gemini app later this year! 🔍 𝗔𝗜 𝗢𝘃𝗲𝗿𝘃𝗶𝗲𝘄𝘀 𝗶𝗻 𝗚𝗼𝗼𝗴𝗹𝗲 𝗦𝗲𝗮𝗿𝗰𝗵: Now available in the US, this tool provides comprehensive AI-generated overviews in your search results. More countries to follow soon! 🎨 𝗜𝗺𝗮𝗴𝗲𝗻 𝟯: This model takes image generation to the next level with more photorealistic creations and improved text rendering capabilities. 📹 𝗩𝗲𝗼: A generative AI tool that produces high-quality 1080P videos through VideoFX, set to compete with OpenAI's Sora. And more exciting products have been announced! For more details, check out the full keynote. At Deloitte, our dedication to innovation ensures we stay at the forefront of technology, delivering tailored, cutting-edge solutions for our clients. #AI #GoogleIO #Innovation #Technology #GenerativeAI #ProjectAstra #Gemini #GoogleAI Alexandre Pinot, Julien Varin, Gilles Lucien, Samy Ait Aissa, Arnaud Guilhabert, Karl-Eduard Berger, Ph. D, Hugo Kisebwe, Shani Dahan

🚀 9 AI Updates from Google I/O 2024! 🚀 Google I/O 2024 was packed with groundbreaking announcements, especially in AI! Here are the top highlights: Bard is now Gemini! 🔹 Gemini 1.5 Flash The latest and fastest large language model (LLM) available in public preview. It offers cost-efficient, high-performance AI capabilities. Gemini 1.5 Pro has also been improved. 🔹 Google Search Major AI enhancements are coming to Google Search! The Gemini integration allows faster results, and the AI overview feature provides conversational summaries of search results. Additionally, AI-organized search pages will offer unique headlines for better search result organization. 🔹 Veo (Text-to-Video Generator) Google’s new text-to-video AI model can generate high-quality 1080p videos longer than a minute! It understands natural language better, allowing users to create different video styles that align more closely with their vision. 🔹 Imagen 3 Google's latest text-to-image generator produces highly detailed and realistic images. It also excels at rendering text within images, addressing a common challenge for AI image generators. 🔹 SynthID Updates Google's SynthID, a technology for watermarking AI-generated images, now supports text and video. This expansion ensures content authenticity and traceability, with all videos generated by Veo including SynthID watermarks. 🔹 Ask Photos A new feature in Google Photos, Ask Photos uses AI to help users find specific images through conversational prompts using Gemini. For instance, you can ask Google Photos to show your daughter's swimming progress over time, and it will automatically compile the relevant highlights. 🔹 Gemini Advanced Upgrades Gemini Advanced subscribers will benefit from several enhancements, including access to Gemini 1.5 Pro with a massive one million token context window. Gemini Live, a new mobile experience, allows real-time voice interactions with natural-sounding voices and the ability to interrupt and use the camera for context. 🔹 AI in Android Android is getting several AI-driven features. “Circle to Search” lets users circle images, videos, and text on their screen to perform a Google search, helping with tasks like solving math and physics problems. Gemini will replace Google Assistant as the default AI assistant. 🔹 Gemini for Google Workspace Updates Google Workspace integrates Gemini 1.5 Pro into its side panel for Gmail, Docs, Drive, Slides, and Sheets. This upgrade allows for more advanced reasoning and longer context windows. Gmail will have features like email thread summarization, Q&A, and Contextual Smart Reply to enhance productivity and user experience. These innovations are set to transform how we interact with technology. Dive deeper into these advancements and see how they can revolutionize your tech experience! Read more here: https://shorturl.at/rGRV6 #GoogleIO #AI #Innovation #TechUpdates #GoogleGemini #Artificialintelligence #GoogleWorkspace

Such a great customer story with #Backcountry. Great nuggets in this story such as "Big Query and Looker have become the brain of our business" and "because of our rock-solid data foundation in BigQuery and Looker, we can now pick and choose the best AI use cases for our business". Putting in the work now to build for the future is critical. #GoogleCloud #AI

Buckle up and get ready to have your mind blown 🤯 with all the new #AI capabilities unleashed at Google I/O powered by #Gemini. Here are some highlights and we see so much opportunity ahead, for creators, for developers, for startups, for everyone. Proud of the tireless efforts by multiple teams to enable these launches. - Multi-modality with 2M context window - AI Agents and Project Astra - Music AI Sandbox Making Loops From Prompts (As a musician, super excited about this!!) - Google Search with AI Overviews and video search! - Workspace side panel powered by Gemini 1.5 Pro - Ask Photos can help you search your memories in a deeper way - 6th generation of TPUs (Trillium) - #Veo, AI video generator #GoogleIO2024 #GoogleIO #GenAI #GenerativeAI https://lnkd.in/g2ps4HZU

Highlighting the important point: Does UX belong in machine learning and AI? It’s such a technical space why would UX even be needed. This is why. When you democratize AI you need ways for people to adopt it and create with it and use it in their everyday lives! That’s what UX does and thanks to AI as the stack shrinks between the technology and the end users UX will be even more important!

ICYMI: This week has been a busy one for big AI tech, with OpenAI announcing GPT-4o and Google sharing their new and improved Gemini 1.5 at the I/O 2024 conference, and we're only halfway through the week. With the introduction of multimodal reasoning and long context windows, and an emphasis on speed and efficiency, both of these tools bring the #artificialintelligence revolution one step closer to an AI agent-powered future. Gemini is free to use as before, however it's notable that GPT-4o is now available for free - a departure from OpenAI's previous subscription-based model. This is an interesting move by OpenAI, possibly an attempt to shorten the gap caused by Meta releasing Llama 3 for free? Or perhaps in beta-testing for it's highly anticipated GPT-5; a form of distraction to buy more time for the next release. Nonetheless, OpenAI is playing chess, not checkers by cheekily announcing GPT-4o a day before Google's I/O conference. Another interesting point to note is that both GPT-4o and Gemini's conversational features are not currently available to the public. Both companies have been in talks with Apple to integrate their AI into iPhones for a much-needed improvement to Siri and AI search capabilities. From a marketing standpoint, an AI-capable iPhone with improved Siri could be just the push Apple needs to bolster the value of it's iPhones, whose global sales have faltered in the last few quarters. Every move in the AI race has been incredibly impressive and made at a ridiculously fast pace. 2024 already seems to be the year of AI agents. What are your thoughts on this? #artificialintelligence #openai #google #googlegemini #aiagent #marketing #apple #gpt4o