The Week in AI, March 4-10: Anthropic, Salesforce, MLCommons, Stability, Inflection, C3 AI, More -- Pure AI

The Week in AI, March 4-10: Anthropic, Salesforce, MLCommons, Stability, Inflection, C3 AI, More

By Pure AI Editors
03/11/2024

This edition of our weekly roundup of AI product and services news includes Anthropic's Claude upgrade, Salesforce's Einstein 1 Studio launch, C3 AI's 2nd edition GenAI, developments in ML Common's Croissant project, Inflection AI's new LLM, Netlify's AI-enabled deployment assistance tool, and more.

Amazon- and Google-backed AI startup Anthropic announced an upgrade of its Claude family of foundational AI models. Claude 3, available via the Google Vertex AI platform, includes three state-of-the-art models: Claude 3 Haiku, Claude 3 Sonnet, and Claude 3 Opus. Haiku focuses on speed and efficiency, ideal for quick customer interactions and content moderation. Sonnet offers a balance between speed and intelligence, suited for enterprise workloads. Opus outperforms rival models GPT-4 from OpenAI and Gemini 1.0 Ultra from Google on various benchmark exams, according to the company. (See our reporting here.)

Enterprise AI software provider C3 AI announced the availability on Google Cloud Marketplace of C3 Generative AI: Standard Edition, its no-code, self-service generative AI application. This release supports Google’s latest large language model, Gemini, and provides simple, self-service, no-code onboarding flow designed to complete in minutes. It enables enterprise users to rapidly locate, retrieve, and act on insights contained in disparate enterprise data via an intuitive search and chat interface. The Standard Edition on Google Cloud Marketplace is currently available for three industry-specific modules: manufacturing, customer service, and financial services. The company is offering a 14-day free trial.

Artificial Intelligence engineering consortium MLCommons announced the release of Croissant, a metadata format developed to help standardize machine learning datasets. According to the ML Commons' Croissant Working Group webpage, Croissant was developed to make datasets easily discoverable and usable across tools and platforms. The Croissant vocabulary is an extension to schema.org, a machine-readable standard for describing structured data, used by more than 40M datasets on the Web, according to MLCommons, which allows the datasets to be discoverable through dataset search engines such as Google Dataset Search. This release includes the format documentation, an open source library, and visual editor, with industry support from HuggingFace, Google Dataset Search, Kaggle, and OpenML, among others.

AI startup Inflection AI launched Inflection-2.5, an upgraded version of its Inflection-2 large language model (LLM), announced last year. The upgrade was designed to match GPT-4’s performance while using only 40% of the computational resources, the company said. This version of its flagship LLM integrates real-time web search capabilities for up-to-date information retrieval. Where the company's Pi "personal AI" was designed to be empathetic ("emotionally intelligent"), Inflection-2.5 enhances logical reasoning and problem-solving, promoting utility in diverse tasks. "Now we are adding IQ to Pi’s exceptional EQ," the company said in a statement.

Cloud web development company Netlify announced the general availability of an AI-enabled deployment assistance tool available as part of Netlify’s Composable Web Platform. The tool is designed to leverage AI to analyze failed deployments and to provide suggestions to correct any errors that are discovered. Using the tool, developers can significantly reduce the time spent manually reviewing logs for increased productivity and faster go-to-market workflows, the company says. Features include the ability to debug failed builds and provide suggested fixes, eliminating the need for "log-diving" and troubleshooting to enable faster code shipping. Doing so, the company says, leads to "a differentiated and more rewarding developer experience."

Cloud computing and social enterprise software-as-a-service (SaaS) provider Salesforceannounced the availability of Einstein 1 Studio, a set of low-code tools designed to enable developers and administrators to customize the company’s AI assistant, Einstein Copilot, and to embed it across any customer management app or AI "experience." Einstein Copilot was designed to enables users to generate responses using their own private and trusted data, while maintaining strict data governance and without requiring expensive AI model training, the company said. The company announced Einstein 1 Studio during TrailblazerDX, the company’s annual developer conference.

HP's PC and printer group, HP Inc. announced a new portfolio of AI PCs, including the new AI HP Elite and Pro PCs, featuring Copilot in Windows7 and Windows 11 Pro8, and Z by HP mobile workstations. The newest generation of AI HP ZBook mobile workstations include a broad range of CPUs including Intel Core Ultra 5, 7 and 9 processors or next-generation AMD Ryzen PRO processors with dedicated NPUs. "At HP, our mission is to empower companies to capitalize on the power of AI, sparking ambitious and meaningful progress in every organization," said Alex Cho, President of Personal Systems at HP Inc., in a statement.

Leading open-source generative AI company Stability.ai, in partnership with AI-power 3D modeling tool maker Tripo AI, released TripoSR, a fast 3D object reconstruction model inspired by the recent work of LRM: Large Reconstruction Model For Single Image to 3D. This new image-to-3D model is designed to cater to the growing demands of entertainment, gaming, industrial design, and architecture professionals, with responsive outputs for visualizing detailed 3D objects, the company said in a statement. The model weights and source code are available for download here under the MIT license, allowing commercialized, personal, and research use.

Cloud content management company Box Inc. announced that it is integrating its Box AI offering with Microsoft’s Azure OpenAI Service. The purpose of this integration is to enable OpenAI’s most advanced large language models to tap into the vast amounts of data that organizations store within the Box Content Cloud. The announcement builds on last year’s launch of Box AI, which introduced a suite of generative artificial intelligence capabilities within Box. Box AI, which was in beta until last week, is focused on helping users ask and answer questions about the information stored within Box Content Cloud. It also helps them tap the generative aspect of generative AI to create new content.