The Week in AI: Oracle, Apple, Meta, xAI, Amazon, Cerebras, Knotch, Ithaka S+R and More -- Pure AI

The Week in AI: Oracle, Apple, Meta, xAI, Amazon, Cerebras, Knotch, Ithaka S+R and More

By Pure AI Editors
03/17/2024

This edition of our weekly roundup of AI product and services news includes Oracle's roll out of a range of GenAI tools for its Fusion Cloud App Suite, Apple's acquisition of DarwinAI, Meta's new 24k GPU clusters, xAI's open-source plan, Cerebras Systems' massive new AI processor, and more.

Oracle rolled out new GenAI capabilities for its Fusion Cloud Application Suite, enhancing areas such as finance, supply chain, HR, and customer experience. The update brings GenAI features to business workflows across multiple domains, including sales and marketing. Additionally, Oracle has expanded its Guided Journeys framework, allowing customization of GenAI features to meet specific industry needs. Utilizing Oracle Cloud Infrastructure, the company offers more 50 generative AI use cases within its applications while prioritizing data privacy and security. The OCI Generative AI Service ensures that customer data is not shared with large language model providers or other customers, and only the data owner can access custom-trained models.

Apple Inc. has acquired DarwinAI, a Canadian artificial intelligence startup, according to Bloomberg. The Cupertino, California-based tech giant finalized the purchase of the explainable AI (XAI) firm earlier this year, with several DarwinAI employees now part of Apple's AI division, reported Mark Gurman. DarwinAI specializes in AI-driven technology that scrutinizes manufacturing components. Its platform, embraced by numerous Fortune 500 companies, also focuses on refining AI systems for enhanced speed and compactness. These advancements align with Apple's strategy of deploying AI directly on its devices, rather than relying solely on cloud-based computing. Alexander Wong, a University of Waterloo AI researcher and DarwinAI co-founder, has been appointed a director within Apple's AI team, as part of the acquisition.

Meta introduced two new 24k GPU clusters designed to support Meta's "current and next generation AI models, including Llama 3, the successor to Llama 2, our publicly released LLM, as well as AI research and development across GenAI and other areas," the company said in a statement. Both clusters are built on the company's Grand Teton hardware, an AI training platform it contributed to the Open Compute Project (OCP). Grand Teton utilizes the company's Open Rack power and rack architecture (also contributed to the OCP), and it uses PyTorch, an open-source machine learning framework based on the Python programming language and the Torch library, to power AI workloads.

xAI, the AI company founded by Elon Musk, is releasing to open source the base model weights and network architecture of Grok-1, its large language model (LLM). Musk made the announcement himself on his X (formerly Twitter) social media account on March 11: "This week, xAI will open source Grok." Grok-1 is a 314 billion parameter Mixture-of-Experts model trained from scratch by xAI. Specifically, the company is releasing the raw base model checkpoint from the Grok-1 pre-training phase, which concluded in October 2023. This means, the company says, that "the model is not fine-tuned for any specific application, such as dialogue." The Grok-1 LLM is being open sourced under the Apache 2.0 license.

AI startup and Nvidia competitor Cerebras Systems unveiled its massive new AI processor, the Wafer Scale Engine 3 (WSE-3). The WSE-3 comprises 900,000 AI cores on a silicon wafer measuring 8.5 by 8.5 inches (21.5 by 21.5 centimeters). Each core is independently programmable and optimized for the tensor-based, sparse linear algebra operations that underpin neural network training and inference for deep learning. The WSE-3 designers spread 44GB of super-fast on-chip SRAM evenly across the entire surface of the process, which gives every core single-clock-cycle access to fast memory at extremely high bandwidth: 21 PB/s. The WSE-3 is set to power the 8-exaFLOP Condor Galaxy 3 supercomputer.

SAP introduced new AI features in its SAP Datasphere and SAP Analytics Cloud platforms, aiming to advance its data management capabilities. The company announced an expanded partnership with data governance firm Collibra as part of the rollout, which SAP CTO Jurgen Mueller described as a "quantum leap" for SAP's support of intelligent business transformation. SAP's updated roadmap introduces "a vital semantic layer to enterprise data," said IDC analyst Robert Parker in a statement, adding that it establishes " the groundwork for AI-driven applications." Launched a year ago, SAP Datasphere is an extensive data service built on top of the SAP Business Technology Platform. It offers a single experience for data integration, cataloging, semantic modeling, warehousing, federation, and virtualization.

Amazon sellers got a new e-commerce platform that enables them to create product listings using GenAI from existing URLs. This enhancement streamlines the listing process, allowing sellers to generate Amazon listings with information pulled from an item's URL on external websites. Initially introduced last year to create listings from minimal text input or images, the tool uses AI to make it possible for sellers to produce high-quality content for Amazon's marketplace with just a URL. The feature is being released to U.S. sellers in English. The company's blog notes the AI tool's capacity to improve engagement, with past data showing listings with contextual backgrounds garnering 40% more clicks than those with white backgrounds.

Ithaka S+R, a not-for-profit service focused on helping academic and cultural communities "serve the public good and navigate economic, technological, and demographic change," announced that it is releasing its GenAI Product Tracker to the public. The Product Tracker was designed to list GenAI products that are either marketed specifically to postsecondary faculty or students, or appear to be actively in use by postsecondary faculty or students for teaching, learning, and/or research activities. "The Tracker is a living document," the company says, "which we update regularly as new products enter the market or new information about existing products becomes available." The company also released an Issue Brief that explores "what these products suggest about the future of generative AI in teaching, learning, and academic research."

Digital intelligence platform Knotch has launched Knotch AIQ, a GenAI solution designed to help users create new content, optimize existing content, and drive performance. The platform draws on Knotch’s audience and performance data to content that will deliver positive results, the company says. Knotch AIQ creates a client-specific performance profile for each brand, which unlocks such data-driven insights as what types of content performs are best for a target audience. Users of AIQ can also leverage GenAI capabilities to create content, and to improve content that hasn’t performed well. The platform also automates labor-intensive manual tasks, such as content tagging and site audits, and it provides predictive insights about the future performance of content.