Day 1 of Microsoft Build 2025 was, as expected, an AI-focused event with Satya Nadella devoting much of the keynote to artificial intelligence agents giving us a picture of a future in  which these agents perform tasks and make decisions on behalf of users and organization. 

First in the spotlight was GitHub Copilot, which according to Nadella is already used by 15 million developers and the fact that it  now includes an asynchronous coding agent, embedded directly in GitHub and accessible from VS Code. 

Previously codenamed Project Padawan, this coding agent was first announced in preview in February for VS Code Insiders and, as we reported, was then rolled out in agent mode in Visual Studio Code in April. Operating within GitHub’s native control layer and powered by GitHub Actions, the agent can test, iterate, and refine code to which developers can delegate routine and specialized tasks.Referring to the way in which the coding agent is intended to collaborate with human developers as a team, Nadella described it as a "peer programmer for the complete software lifecycle.

Nadella also announced that GitHub Copilot Chat is being open sourced in Visual Studio Code, reinforcing Microsoft's commitment to open collaborative AI powered software development. 

GitHub Copilot wasn't the only coding agent to be in the limelight. Sam Altman joined Nadella by video link and talked about Codex, OpenAI's tool designed to generate code, fix issues, and write tests - all with minimal input from users. It operates on the codex-1 model, a variant of OpenAI’s o3 model. Like Copilot, Codex is  becoming an agent that works alongside developers with OpenAi claiming:

“Eventually, Codex will feel less like a tool and more like a remote teammate you collaborate with asynchronously"

Later it was Elon Musk's turn to appear in a video call to talk about Grok, the family of large language models developed by xAI to push the boundaries of AI innovation and accelerate human scientific discovery.

Unlike other AI models which only have knowledge relating to the past, Grok has direct access to real-time data, particularly from X (formerly Twitter) which allows it to provide up-to-date responses on current events and trending topics. According to xAI Grok 3, its flagship model, blends unparalleled reasoning with vast pretraining knowledge, honed on xAI’s Colossus supercluster with 10x the compute power of prior leading models and that it  excels in reasoning, mathematics, coding, world knowledge, and instruction-following, setting a new standard for enterprise AI. 

The announcements at Build were that Grok 3 is now available in GitHub Models and in the Azure AI Foundry models for a free preview for the next 2 weeks. It is suitable for tasks such as data extraction, coding, and text summarization in domains like finance, healthcare, law, and science, supporting knowledge-driven workflows and professional use cases.   

More Information

GitHub Introduces Coding Agent For GitHub Copilot

Announcing Grok 3 on Azure AI Foundry

Related Articles

GitHub Copilot Adds VSCode Agent Mode

Introducing Codex

GitHub Copilot Gets Agent Mode

To be informed about new articles on I Programmer, sign up for our weekly newsletter, subscribe to the RSS feed and follow us on Twitter, Facebook or Linkedin.

Comments

Make a Comment or View Existing Comments Using Disqus

or email your comment to: [email protected]