Redmond: Microsoft is considering a future where agents of artificial intelligence of a company can work with agents of other companies and having better memories of its interactions, said its chief technologist on Sunday before the annual conference of software developers of the company.
Microsoft is organizing its construction conference in Seattle on May 19, where analysts expect the company to reveal its latest tools for AI systems creation developers.
Speaking at Microsoft’s headquarters in Redmond, Washington, before the conference, the director of technology, Kevin Scott, told journalists and analysts that the company focuses on the stimulation of the adoption of standards in the technological industry which will allow agents of different manufacturers to collaborate. Agents are AI systems that can accomplish specific tasks, such as the correction of a software bug, by themselves.
Scott said Microsoft supports a technology called Model Context Protocol (MCP), an open source protocol introduced by Anthropic supported by Google. Scott said MCP has the potential to create an “agentic web” similar to the way the hypertext protocols that helped distribute the Internet in the 1990s.
“This means that your imagination can lead what the agentic web becomes, not just a handful of companies that have first seen some of these problems,” said Scott.
Scott has also said that Microsoft is trying to help AI agents to have better memories of things that users have asked them to do, noting that, so far, “most of what we build feels very transaction”.
But making the memory of an AI agent better costs a lot of money because it requires more computing power. Microsoft focuses on a new approach called structured recovery increase, where an agent extracts short pieces from each turn in a conversation with a user, creating a roadmap for what has been discussed.
“This is an essential part of how you train a biological brain – you don’t force raw in your head whenever you need to solve a particular problem,” said Scott.
