Azure Updates: AI supercomputing; Data Share & Data Explorer; Synapse Analytics; Blob Storage

Over the past week, Microsoft emphasized new storage, data and AI supercomputing capabilities while also casting its attention to IoT Central and other Azure services.

Microsoft senior program manager Ian Finder discussed the ways that the company is working to bring AI supercomputing to customers. The Azure team's Turing Natural Language Generation model relies on 17 billion parameters to summarize documents, assess language data and move beyond the limits of older image-centric models—a process that involves hundreds of machines running AI accelerators. Back in May, OpenAI unveiled a 175 billion parameter GPT-3 model. Currently, Microsoft is developing new capabilities around Azure Machine Learning, ONNX Runtime and Azure AI services to meet its supercomputing goals. With similar goals in mind, Ricky Kapur, VP of Sales, Marketing and Operations for Microsoft Asia Pacific announced a Virtual Hackathon for that region focused on data science and drone-related development.

Principal program manager Manoj Raheja encouraged users to share data at scale using Data Share and Data Explorer, both of which are GA and provide close to real time collaboration as an alternative to slower data sharing with File Transfer Protocols, APIs and attachments. Once a sharing relationship is established, data consumers can read and query data through the Data Explorer cluster.