Two years later GitHub VP sees continued growth, simplification, AI explosion

iTWire spoke with Sharryn two years ago when she first took the reigns in APAC. At that time, her appointment was intended to accelerate momentum in the APAC region for the open source and software developer hub.

And what a two years it has been; in this time the region has grown from 960,000 developers using the platform to over 1.4 million, with 20% growth year-on-year.

GitHub, of course, is the largest source code repository on the Internet, with millions upon millions of projects, authors, and petabytes of code. It provides free source code hosting, as well as enterprise features for teams, and caters for both public open source programs as well as private repositories. It's in a remarkable position where anything GitHub does literally affects developer productivity worldwide, whether good or bad. If GitHub can shave a second off an interaction here or there, this magnifies to huge time savings globally. Conversely, if GitHub damages a workflow, then that similarly ripples to the corners of the Earth.

Ideally, the company is providing the first type of change, and these kinds of changes are something Sharryn Napier says is resonating well with large enterprises. "Over the last two years we've seen a lot of simplification and tool consolidation of large enterprises," she said. "It's something GitHub can solve for big organisations, particularly banks, who are leaning in because of our AI-driven developer platform."

"Getting rid of noise and other point solutions brings communities together inside an organisation," she says. "And, taking on GitHub allows them to integrate better, and be more productive and share more together."

"As we've moved towards a full AI-driven platform we're going to find additional benefits beyond productivity, that will be enormous."

Anecdotally, customers have reported to Napier that they've seen improvements in code quality and build times through GitHub. These conversations were developer-focused, and while these reports continue to be positive, it's a "super different story now," she says, with an increasing number of conversations coming from the business side of large organisations.

In the past, it was unheard of for a bank executive to come out to GitHub and talk about developer tools. "But now, CBA, Westpac, NAB ... they're all talking about GitHub Copilot," she said.

Over 50,000 organisations are now using Copilot, Napier advised - an especially impressive number when you consider the product was only released to the enterprise market in February last year. "We're 12 months in, and to have the impact on these organisations, 1.3 million people, it's a bit of an indication of where it is going to go," she said.

This AI-infused approach is something GitHub is increasingly offering, with Copilot that provides intelligent, predictive assistance in writing code itself, through to identifying credentials and shared secrets inadvertently exposed in code, to automatically remedying potential vulnerabilities coming from dependencies.

"The whole platform AI-infused approach will see organisations create pull requests, raise issues, fix code vulnerabilities automagically," Napier said. "Companies, particularly banks, are seeing extended value in these AI offerings beyond Copilot itself."

It's one thing to make the claim AI is helping people write code faster and better, reducing vulnerabilities, and another thing to prove it. "We have the emperical evidence," Napier said.

GitHub has previously released figures that found its AI tools have helped a whopping 87% of developers preserve mental energy, and 74% experienced an increase in overall daily workflow satisfaction. Code is being produced 55% faster, and 85% of developers feel more confident over their code quality.

GitHub saw a lot of people looking at GitHub Copilot and saying "let's prove what we can do," Napier says. "We saw a lot of experimenting and a lot of emperical evidence, moving into full-blown use cases in generative AI and Copilot."

"A lot of organisations have now gone 'tick' - it gives us greater productivity, greater code quality, improved build times, and now identifies and autofixes vulnerabilities."

"These organisations are now saying 'how do we really use this?'," she said. "In the next 12 months, we're going to see more practical applications of all this work."

For example, Napier explains new developers can come in and instead of taking months to get familiar with legacy code they can simply ask the chatbot "What does this code do?". Or, they can summarise the backlog. All kinds of things become possible when you have assistive AI tools, that are proven and trustworthy.

"The next 12 months will be the practical evidence, with companies touting what they will have achieved, not simply testing and evaluating."

In addition, Napier is involved in the Code without Barriers initiative. This was born in Microsoft in Asia and is now being deployed throughout the rest of Asia and Australia and New Zealand.

"The initiative reduces the barrier to women in tech," Napier explained, and this offers another lens where we see the benefits of generative AI.

For instance, a lone female coder in a team of men might feel intimidated and reluctant to ask questions. "Copilot never says 'that's a dumb question'," Napier said. "It will always help people learn, and can lower the barrier to entry. GitHub's AI tools offer several aspects beyond productivity benefits alone, and opens us up to bring in and enable more people in the tech platform and environment."

Another example Napier offers is from one of her big bank clients. "It dawned on them they can ask Copilot to explain legacy code - old COBOL code that was never documented," she said.

Code like this can be scary; developers - many of whom may not have been born when the code was first written - don't want to break it. "It runs ok, so they adopt the mantra 'don't touch'," she said. "AI allows us a different dynamic."

Whether it's new projects, understanding historically frightening legacy code, onboarding new team members, and reducing gender imbalances, "all these different pieces can be peeled back with an AI-infused platform like GitHub," Napier said. "It's super exciting."

"Australian financial services like ANZ and CBA can really showcase to the world what innovation and creativity can do for them."

So what else is coming up on the GitHub roadmap? "It's all about using generative AI in practical use now," Napier said. "We will see different business models around AI services, different things, new business models. We will see interesting things created in the next 12 months to two years that we never thought about before."

"It's a time of innovation and GitHub is at the forefront and leading," she said. "We see, lead, and project what people are doing, and help them make solutions."

"The next few years will be incredible for this whole space," she said.

"As we continue to infuse AI through GitHub, the benefits will be exponential. Things will be done more automatically, and we won't have to do things that developers hate like pull request summaries. The infusion in the platform will help them - not just developers right now, but the people coming behind them."

"It's a great thing to see as we move into the GitHub AI platform."

BONUS DATA: Facts and figures on GitHub's AI products

GitHub Copilot

• Over 1.3 million paid GitHub Copilot subscribers and over 50,000 organisations across all geographies and sectors using the tool.
• Developers are writing code over 55% faster with Copilot.
• Last year, GitHub saw 35% of newly written code was suggested by Copilot. Now, in files where Copilot is enabled, up to 60% of the code is being written by Copilot in popular coding languages like Java.
• In the next five years, GitHub expects this to grow up to 80%.
• We recently did an experiment where participants authored and reviewed code using Copilot and Copilot Chat, and found that 85% felt more confident in their code quality, code reviews were completed 15% faster, and 88% felt less frustrated and more focused when coding.

Australian stats from the 2023 Octoverse report

• Over 1.4M developers on GitHub in Australia
• 21% increase in developer growth on GitHub in Australia in 2023

ANZ Bank

• GitHub Copilot has steered software engineers at ANZ Bank toward improved productivity and code quality.
• From mid-June – July 2023, ANZ Bank conducted an internal trial of GitHub Copilot that involved 100 of the bank’s 5,000 engineers.
• The group that had access to GitHub Copilot completed their tasks 42% faster than the control group participants.
• This research provides compelling evidence of the transformative impact of GitHub
• The adoption of this tool has marked a shift, empowering engineers to focus more on creative and design tasks while reducing time spent on repetitive boilerplate tasks.
• GitHub Copilot has now been widely adopted within the organisation, with over 1,000 developers using it in their workflows.

CBA

• The bank’s team was among the first in Australia to test GitHub Copilot, seeking to deliver code faster, protecting the bank and getting more done for its customers.
• CBA began testing the service with 200 people and discovered that 75% of its engineers found it “very helpful”. They accepted nearly 80,000 lines of code that GitHub Copilot recommended – a third of all the recommendations made.