An anonymous reader shares a report: Since 2017, Microsoft has pursued this goal under the name Project Turing, a team that’s tasked with building these large language models and figuring out how they can be used in the company’s vast suite of products. Project Turing might not be a visible name outside the company, its AI can already be found generating text inside Microsoft Office products and powering much of the curated information provided when searching with Bing. If Turing succeeds, the strategy could amplify the research dollars that Microsoft has poured into AI research over previous decades. Notably, Microsoft isn’t only using Turing-NLG, the project’s flagship model, internally: It’s already begun selling the tech to select partners, hinting at the cloud giant’s ambitions for the AI market. Insider spoke with AvePoint and Volume.ai, both of whom are using Turing in their own products.
“Our job is to further the frontier of AI innovation as much as possible,” Ali Alvi, group program manager of Project Turing, told Insider. Alvi tells Insider that the Turing team was assembled from within the company by Microsoft chief technology officer Kevin Scott, in recognition of the ongoing deep learning boom. Scott encouraged the team to think bigger and work with the Azure infrastructure team to make the models exponentially larger. When CEO Satya Nadella saw the team’s progress, he decided to get it into the hands of customers, Alvi says. AvePoint, a Microsoft partner that resells and builds applications on top of Microsoft products, has launched two products so far using the Turing model: An education platform for teachers that will automatically create quiz questions using material that’s been uploaded for a specific course, and a corporate training platform that uses Turing to test employees on internal material.
of this story at Slashdot.
Source:: Slashdot