At GitHub Universe, GitHub unveiled a new era of multi-model choice for Copilot, introducing Anthropic’s Claude 3.5 Sonnet, Google’s Gemini 1.5 Pro, and OpenAI’s o1-preview and o1-mini, empowering developers and enterprises to select models tailored to specific coding needs across tools like VS Code and GitHub.com.
“The next phase of AI code generation will not only be defined by multi-model functionality but by multi-model choice,” said GitHub, underscoring GitHub’s commitment to open developer choice. Alongside this, GitHub previewed “Spark,” a natural-language tool to create AI-integrated applications, supporting GitHub’s vision to reach 1 billion developers.
“In 2024, we experienced a boom in high-quality large and small language models that each individually excel at different programming tasks. There is no one model to rule every scenario, And developers expect the agency to build with the models that work best for them,” said GitHub CEO Thomas Dohmke at the GitHub Universe event.
Anthropic’s chief science officer Jared Kaplan took the stage, sharing insights on Claude 3.5’s exceptional appeal: “Our mission has always been to get the most frontier model capabilities in as many people’s hands as quickly as possible. If we can’t deploy them now, they’ll be obsolete in six months to a year.”
Further broadening their multi-model strategy, GitHub revealed another major integration with Google’s Gemini 1.5 Pro, a model known for its “two-million-token context window and native multimodality,” designed to process “code, images, video, and text simultaneously,” as GitHub demonstrated live.
In an innovative Copilot update for VS Code, GitHub’s senior director of developer advocacy Cassidy Williams, showcased Copilot’s expanded capabilities, including custom instructions, multi-file editing, and integration with GitHub’s search and intent detection tools, giving users control over how they code.
“This is your brand-new experience for VS Code,” Cassidy noted, emphasising that the latest features—model selection, repo indexing, and GitHub Copilot extensions—are “available to every single Copilot user, from individuals to enterprises.” starting from this week.
For the first time, GitHub Copilot will also extend beyond VS Code, with support announced for Apple’s Xcode, enhancing accessibility for mobile and desktop developers alike, regardless of their preferred platform or workspace.
More Choice & Agency to Developers
Moving on from Codex. Earlier this month, Microsoft’s GitHub introduced OpenAI’s o1-preview and o1-mini models on Azure. These models are available through GitHub Copilot and Models, and developers can sign up to access OpenAI o1 in GitHub Copilot Chat via VS Code and the GitHub Models playground.
Now, the integration of Claude 3.5 Sonnet with Github is live. “Claude 3.5 Sonnet excels at coding tasks and is broadly used by developers for its exceptional grasp of software engineering principles and ability to tackle complex programming challenges,” said Jared Kaplan, co-founder and chief scientist at Anthropic.
Google’s Gemini 1.5 Pro’s integration with GitHub will be available in a few weeks. Developers are seeking different models for tasks like code generation, refactoring, and optimisation, enabling flexible, efficient, and high-quality coding across programming environments. “Gemini models excel at this and are accessible on widely used developer platforms and environments – including now with GitHub Copilot – so millions of developers globally can benefit from trusted, enterprise-grade AI through Google Cloud,” said Thomas Kurian, CEO at Google Cloud.
Kills Cursor’s Vibes
In a recent podcast with Lex Fridman, Cursor’s co-founders highlighted how Microsoft-backed GitHub Copilot is falling behind startups in terms of innovation. They emphasised how the tech giant lacked the research and experimentation necessary to really push the ceiling.
The sentiment also remains that Copilot did not have any ‘alpha features’ for a very long time. Even Y Combinator funded several open-source AI code editors like Continue, Pearl, Void, Type, and Melty, among others.
The Copilot experience is finally improving with substantial updates to GitHub Copilot in VS Code, Copilot Workspace, GitHub Models, and Copilot Autofix. We have written extensively about whether it is too late for Microsoft’s VS Code or GitHub Copilot to catch up with the market.
Now the tables have turned. The days may be numbered for Cursor and other AI coding assistants, as GitHub Copilot advances toward becoming the go-to cross-platform solution for developers.
“It is clear the next phase of AI code generation will not only be defined by multi-model functionality, but by multi-model choice. Today, we deliver just that,” said Dohmke.
With this, developers can choose the right model for the right use case or continue to let Copilot use its powerful default. GitHub is allowing developers to build with an array of leading models in the workflows they’re accustomed to.
Aman Sanger, co-founder of Anysphere (the creator of Cursor), said, “I think the Cursor a year from now will need to make the Cursor of today look obsolete.” This indicates if Cursor is to stay relevant in the competition, it has to pull its socks up.
“You can wax poetic about moats and brand that and this is our advantage, but I think in the end, just if you stop innovating on the product, you will lose,” said Micheal T, a co-founder of Anysphere. It’ll be interesting to see how Cursor responds, and how the competition will evolve.
The post GitHub Kills Cursor Vibes appeared first on AIM.
