Microsoft is spreading OpenAI everywhere to retain software partners

If there’s one company that has popularized artificial intelligence over the past year, it’s the small but richly funded startup OpenAI, the entity behind viral chatbot ChatGPT.

This week at its Build conference for software developers, Microsoft utilized its collaboration with the startup to a large extent, where it has invested billions.

Front and center on Tuesday, the first day of the show, was an onstage conversation between Greg Brockman, OpenAI’s co-founder and president, and Kevin Scott, Microsoft’s chief technology officer and the person credited with building unusually close relationship between the two companies.

“You heard it from Greg,” Scott told the crowd gathered at the Seattle Convention Center near the end of the talk. “All of you are the ones who will make AI great.”

Toward that end, Microsoft announced a series of products for developers that build on OpenAI’s technology:

• There are new ones Azure cloud tools for custom text summary.
• An upcoming chatbot promises to help developers work with data and prepare it for analysis.
• Developers will be able to build plugins that work inside ChatGPT and the chatbots in Microsoft’s own products, including one that will debut in Windows next month.
• Developers who receive coding suggestions via GitHub Copilot the feature will access a chatbot inside the Windows Terminal command line program.

OpenAI released ChatGPT to the wider world in November, attracting a lot of interest from consumers. Shortly afterwards, companies such as Atlasian, Morgan Stanley and Salesforce rushed to show off integrations of OpenAI’s GPT-4 major language model, which powers the chatbot. GPT-4 and alternatives from the likes of Amazon and Google have been trained on vast datasets on the Internet and have become capable of spitting out chunks of natural-sounding text.

The technology is a popular form of what has come to be known as generative AI, which can take human input and respond with a computer-generated output.

“Every layer of the software stack is going to change forever, and there’s no better place to start than the actual developer stack,” Microsoft CEO Satya Nadella said during his Build keynote on Tuesday. “We as developers, the way we build is fundamentally changing.”

It is critical for third-party developers to enrich Microsoft’s proprietary software capabilities, such as the Microsoft 365 productivity suite. Such work, for example, could help Microsoft’s Teams communications app become a more obvious hub for an ever-widening selection of processes and tasks that businesses need to perform. That could make companies less likely to switch to alternatives like Google Workspace.

Microsoft highlighted dozens of plugin developers on Tuesday, including Adobe, Asanacanvas, Cloudflare, Redfine, Spotify and TripAdvisor. A demonstration showed the Windows chatbot activating a Spotify playlist, creating a company logo using Adobe Express, and sending the logo to a person’s colleagues via Teams in response to a series of typed messages.

Meanwhile, Nadella has pushed for Microsoft to incorporate GPT-4 directly into Teams and older Microsoft products, such as the Bing search engine, often resulting in bots named Copilot. The Copilot term emphasizes collaboration with humans, as opposed to (for example) Autopilot’s advanced driver assistance system for Tesla vehicle.

“We’re adding Copilot to everything,” Scott Guthrie, executive vice president of Microsoft’s cloud and AI group, told CNBC in an interview last week. “It’s less of a top-down mandate, although we’re really pushing from the top down. I think it’s something where we’ve actually evangelized internally and really gotten all the teams excited about it. And we’re building a common stack across Microsoft that the whole company is built on top of it.”

Analysts responded positively to the developer attack.

“The pace of MSFT’s GenAI innovation remains fantastic for us,” Mizuho analysts with a buy rating on Microsoft shares wrote in a note to clients on Wednesday.

Brockman hinted to developers that the cost of GPT-4, which runs in Azure, could come down.

“I think we did a 70% price cut two years ago,” he told Scott. “Basically, in the last year, we did a 90% cost reduction. A 10-fold cost reduction—like, that’s crazy, right? And I think we’ll be able to do the same thing repeatedly with new models. And so GPT-4 right now, it’s expensive, it’s not fully available. But that’s one of the things that I think will change.”

LOOK AT: Microsoft Build 2023 reveals plugins and products that incorporate AI