Google kicked off Google Cloud Next ’24 in Las Vegas yesterday and there are some great announcements where we will walk you through. What happened in Vegas on the first day of Next? In one word, AI.
Google began by showcasing how companies are currently leveraging AI, emphasizing its present-day applications rather than viewing it as a future concept.
Google and Alphabet CEO Sundar Pichai then joined the Next keynote on screen, reminding us how far Google has come with generative AI — and how fast. “Last summer, we were just beginning to imagine how this technology could transform businesses, and today, that transformation is well underway,” Sundar said.
During the keynote address, Thomas Kurian and other top executives showcased renowned brands such as Goldman Sachs, Mercedes, Uber, Walmart, and Wayfair, highlighting their innovative use of AI technologies.
During the event, remarkable new products, partnerships, and technological innovations were shown. Here's a brief overview of the major announcements made today, categorized into four essential themes:
Echoing Sundar Pichai's words, the aim is to harness AI to craft products that are significantly more beneficial. Whether you're a developer crafting groundbreaking applications, an architect orchestrating and overseeing infrastructure, an end-user collaborating on projects with your team, or a data scientist delving into your organization's data, the new offerings from Google AI are designed to enhance your work efficiency across various roles.
Among the key announcements:
Gemini for Google Cloud: This innovation is designed to empower users to create AI agents that boost work and coding efficiency. It enables enhanced management of applications, deeper insights into data, and the identification and mitigation of security threats. Gemini is seamlessly integrated with a broad spectrum of Google Cloud services, ensuring users can leverage its capabilities across their cloud infrastructure and applications.
Gemini Code Assist, the evolution of the Duet AI for Developers, lets developers use natural language to add to, change, analyze, and streamline their code, across their private codebases and from their favorite integrated development environments (IDEs).
Gemini Cloud Assist helps cloud teams design, operate, and optimize their application lifecycle.
Continuing the focus on innovation, Google Workspace stands as a testament to the foundational principles of real-time creation and collaboration. In today's announcements, this commitment was elevated further with the introduction of new enhancements and features designed to streamline workflows and foster even more dynamic collaborative environments.
Google Vids: This innovative addition to the Google Workspace ecosystem acts as an AI-powered assistant for video creation, writing, production, and editing. Designed to complement existing productivity tools such as Docs, Sheets, and Slides, Vids aims to empower every user to become an adept storyteller in the workplace, simplifying complex video production tasks into manageable, user-friendly processes.
AI Meetings and Messaging Add-on: A notable enhancement to the Google Workspace suite, this add-on introduces several AI-driven features to revolutionize the way meetings and communications are conducted. Highlights include "Take notes for me" (currently in preview), set to simplify note-taking during meetings, and "Translate for me" (releasing in June), which promises to break down language barriers by providing real-time translation services. Additionally, the add-on offers automatic translation of messages and on-demand conversation summaries in Google Chat, ensuring no detail is missed. Priced at $10 per user, per month, this add-on is designed to integrate seamlessly with most Workspace plans, offering a substantial boost in productivity and collaboration.
You might have guessed it, the core technology enabling these cutting-edge capabilities is Gemini, which stands as Google's most potent AI model to date. At Next '24, the spotlight was also on empowering developers to harness the remarkable capabilities of Gemini within their own applications. This is primarily facilitated through Vertex AI and various other AI development platforms, opening up new avenues for innovation and application enhancement across the board. This initiative aims to democratize access to Google's advanced AI technology, enabling developers to infuse their applications with the intelligence and efficiency that Gemini offers.
What Google announced on this front:
Gemini 1.5 Pro has marked its entry into the public domain, now available in a public preview for Vertex AI customers. This development ushers in a new era for developers, offering them a firsthand experience of crafting solutions with an unprecedented 1M context window. This expansive context window allows for deeper, more nuanced interactions and understanding within AI applications, paving the way for more sophisticated and context-aware AI functionalities.
Google's Imagen has been updated to possess the capability of generating live images directly from text prompts, currently in preview. This innovative feature enables the creation of animated images, such as GIFs, from mere textual descriptions, opening a myriad of creative possibilities. Further enhancing Imagen's functionality are advanced photo editing tools, including inpainting and outpainting, alongside a digital watermarking feature. This watermarking feature is powered by Google DeepMind's SynthID, offering an additional layer of authenticity and security to digital creations.
Vertex AI has also been augmented with new MLOps capabilities, aimed at streamlining and optimizing the machine learning development process. These include Vertex AI Prompt Management and new Evaluation tools, designed to enhance the efficiency and effectiveness of AI model training and evaluation.
Vertex AI Agent Builder signifies a major advancement in the development and deployment of generative AI experiences. By amalgamating Vertex AI Search and Conversation products with an array of enhanced tools, this platform simplifies the process for developers to create and implement enterprise-grade generative AI experiences. These developments collectively signify Google's commitment to providing robust, innovative tools that empower developers to build cutting-edge, AI-driven applications and experiences.
The groundbreaking advancements and innovations presented wouldn't be feasible without the substantial investments in workload-optimized infrastructure. These investments are crucial for powering not only Google's internal systems but also those of its users. This dedication to enhancing infrastructure ensures the delivery of high-performance, efficient, and reliable services, enabling both Google and its customers to harness the full potential of advanced technologies like AI and machine learning. It's this foundational support that underpins the remarkable capabilities and services announced, setting the stage for ongoing innovation and development.
What Google announced:
AI Hypercomputer Architecture Enhancements: The introduction of Cloud TPU v5p and A3 Mega VMs, both of which are designed to offer substantial improvements in processing power for AI and machine learning workloads. The Cloud TPU v5p is now generally available, bringing optimized performance for the most demanding AI tasks. A3 Mega VMs, powered by NVIDIA H100 Tensor Core GPUs, represent a leap in computational capability, tailored for intensive AI applications. Additionally, storage solutions have been optimized with the introduction of Hyperdisk ML, aimed at meeting the high-speed, high-capacity needs of machine learning workflows. Open software advancements are also part of the package, with the release of JetStream and updates to JAX and PyTorch/XLA, enhancing the efficiency and accessibility of AI development tools.
Google Axion - Custom Arm®-based CPU: Marking a significant milestone, Google announced the debut of Google Axion, its first custom Arm®-based CPU specifically designed for data center applications. This development is a clear indication of Google's direction towards more energy-efficient, high-performance computing solutions tailored for the modern data center's needs.
New Compute and Networking Capabilities: The announcement of new general-purpose VMs, C4 and N4, powered by 5th Generation Intel Xeon processors, showcases Google's commitment to providing versatile and powerful computing options. These VMs are engineered to support a wide range of workloads, offering enhanced performance and efficiency. Furthermore, Google has introduced enhancements across its Google Distributed Cloud, aiming to bolster its networking and compute capabilities, thereby ensuring more robust and reliable infrastructure services for users.
At the core of AI's transformative potential is the quality of data it's trained on, a principle well understood at Google Cloud. They champion the concept of "enterprise truth," recognizing the unparalleled value of data housed within trusted enterprise databases and data warehouses. This emphasis on quality and trustworthiness forms the foundation of their approach to AI development. Google Cloud is dedicated to enhancing its Data Cloud capabilities, ensuring that AI applications are built upon a solid foundation of reliable and accurate data. This commitment not only bolsters the performance and effectiveness of AI solutions but also ensures they are rooted in data integrity, offering businesses a competitive edge through informed, data-driven decision-making.
What Google announced:
Databases Enhancements: Google Cloud databases have been upgraded to be more AI-capable. AlloyDB AI now boasts new vector capabilities, providing easier access to remote models and offering flexible support for natural language processing. Firestore has been added to the growing list of Google databases equipped with advanced vector search capabilities, enhancing the ability to handle complex queries and data retrieval tasks efficiently.
Big News for BigQuery: BigQuery is being positioned as the unified data analytics platform within Google Cloud, capable of operating across different cloud environments through BigQuery Omni. This move signifies a major step towards seamless, cross-cloud data analytics. Additionally, BigQuery is introducing a new data canvas, a natural language-based interface designed to simplify data exploration, curation, wrangling, analysis, and visualization processes, making data analytics more accessible and intuitive.
Integration with Google Search for Model Training: In an innovative move, Google Cloud now allows models to be grounded not just in enterprise data but also in Google Search results. This means AI models can be trained with the latest, high-quality information available on the web, vastly expanding the scope and relevance of data sources for more accurate and up-to-date insights.
The announcements made during the morning keynote session highlight Google Cloud's commitment to enhancing the capabilities and accessibility of AI and data analytics technologies. With more than 300 Spotlights and breakout sessions, the Innovators Hive, and a show floor featuring over 100 AI solutions built with Google Cloud technologies, attendees have amazing opportunities to explore the latest in tech innovation.