The new GitHub Copilot helps people code 55% faster, says Shuyin Zhao, of GitHub

CXOToday has engaged in an exclusive interview with Ms. Shuyin Zhao, Sr. Director, Product Management, GitHub.

Q: How is AI transforming the world? GitHub Copilot being an AI developer tool, how does it bring greater benefits to enterprises?

The rise of generative AI over the last several months has sparked industry-wide acceptance that the age of AI is here. In GitHub’s case, over 400+ organizations are already using GitHub Copilot – as of Feb 14 – and this is having a tangible impact on productivity, developer happiness, and the pace at which organizations can innovate.

Since we launched GitHub Copilot, we’ve seen it redefine productivity for more than a million developers – our research shows that GitHub Copilot helps people code 55% faster. When you quantify that organization-wide, it’s so exciting to imagine what that can do for enterprises – for their backlogs, and for developer velocity. Additionally, GitHub Copilot helps developers stay in the flow, focus on more satisfying work, and conserve mental energy, which leads to increased developer happiness. All of this has a direct impact on organizations and their ability to drive innovation.

Q: What does the new update to GitHub Copilot entail?

We’ve made three key upgrades over the past six months for all users for individual and enterprises, including:

• An upgraded AI Codex model: In September 2022, we upgraded GitHub Copilot to a new OpenAI Codex model which delivers better results for code synthesis.
• A lightweight client-side model: In October, we updated the GitHub Copilot extension for VS Code with a lightweight client-side model that improved overall acceptance rates for code suggestions, which as a result reduces the frequency of unwanted suggestions when they might prove disruptive to a developer’s workflow. This resulted in a 4.5% reduction in unwanted suggestions, helping GitHub Copilot better respond to each developer using it. And with a second, improved iteration of this client-side model that we shipped in January 2023, we’ve seen even better improvements in overall code acceptance rates.
• Better context understanding w/ FIM: In December 2022, we improved GitHub Copilot by a new paradigm called Fill-In-the-Middle (FIM) — which offers developers better craft prompts for code suggestions. Instead of only considering the prefix of the code, it also leverages known code suffixes and leaves a gap in the middle for GitHub Copilot to fill. This way, it now has more context about your intended code and how it should align with the rest of your program. FIM in GitHub Copilot consistently produces higher quality code suggestions, and we’ve developed various strategies to deliver it without any added latency.

You can learn more about these updates in our blog here.

Q: How do you see the massive shift towards adopting next-generation AI tools by tech companies?

Industry-wide, we’re now entering a new wave of innovation powered by AI. Ultimately, organizations that harness AI to empower their developers will win, and we recognize this. For us, GitHub Copilot is just the beginning of our vision to build an integrated, AI-powered GitHub that allows developers to build their best software. We plan to integrate AI into every aspect of the developer experience so developers can build their best software.

Q: What are the common issues targeted by AI based security?

We launched an AI-based vulnerability prevention system that blocks insecure coding patterns in real-time to make GitHub Copilot suggestions more secure. Our model targets the most common vulnerable coding patterns, including hardcoded credentials, SQL injections, and path injections.

Q: Please mention the major capabilities of GitHub Copilot that boost developer productivity

GitHub Copilot sits within the editor as a simple extension, and draws context from a developer’s code to suggest new lines, entire functions, tests, and even complex algorithms. GitHub Copilot works with code and natural language prompts to offer multiple suggestions that can quickly be accepted or rejected — and it learns alongside developers to adapt to individual coding styles and conventions. Today, developers all over the world are using GitHub Copilot to code faster, focus on business logic over boilerplate, and this allows them to do what matters most – building great software.