S’pore has 1.3m coders, ranks 9th globally for helping AI projects: GitHub

SAN FRANCISCO – Out of a population of six million in Singapore, more than 1.3 million are coding to solve tech issues.

And that number is growing among the fastest worldwide, at a rate of 28 per cent in the 12 months to September 2024, according to developer platform GitHub.

Boosted by the Government’s efforts to get local workers and students trained in artificial intelligence (AI), Singapore’s coding community also punched above its weight, ranking ninth globally for contributors creating code for AI projects on GitHub in the same period.

More than 9,700 coders in Singapore contributed to AI projects on the open source platform, where code is available for users to view, modify and distribute.

GitHub, in its annual Octoverse report, said: “In 2024, there was a 59 per cent surge in the number of contributions to generative-AI projects on GitHub and a 98 per cent increase in the number of projects overall – and many of those contributions came from places like India, Germany, Japan and Singapore.”

The findings, based on unique GitHub users logging in from Singapore, were released at the GitHub Universe conference in San Francisco, which drew 3,000 participants on Oct 29 and 30.

At the conference, GitHub chief executive Thomas Dohmke and other presenters introduced new AI tools and features, including a headliner which lets developers choose which large language models from Anthropic, Google and OpenAI they want to power their GitHub Copilot, the platform’s AI coding assistant.

“There is no one model to rule every scenario, and developers expect the agency to build with the models that work best for them,” Mr Dohmke said.

Mr Dohmke, a developer himself, also launched the pilot for GitHub Spark, an AI tool that makes web apps with natural language prompts.

He demonstrated the new tool by building a tic-tac-toe game within minutes.

“For far too long, there has been a high, solid, concrete barrier that prevented the vast majority of the world’s citizens from creating software,” he said, explaining the appeal of Spark. With natural language coding, the tech economy will now have fewer limits on opportunities, he added.

GitHub, which has more than 100 million users, will also inch closer to its vision of having one billion users by as early as 2030 with Spark, which comes as more non-English speakers utilise multilingual generative-AI tools for coding.

Ms Sharryn Napier, GitHub’s vice-president for Asia-Pacific, said that even in Singapore, where roughly one in four is a GitHub user, growth continues as more non-tech individuals start coding.

Companies are increasingly seeking AI skills in software-developer hires, she added.

Asked if it could lead to firms hiring more Singaporeans as developers, she said: “That’s one of the things that we see as a big change, the way that the resources are sourced and the way people farm out the projects.”

There are no plans to raise charges for the added tools, said Mr Dohmke.

Individuals pay US$10 (S$13) a month or US$100 annually, while businesses pay US$19 or US$39 a month. GitHub is free for students, teachers and the maintainers of popular open source projects.

Chief of enterprise research at CCS Insight Bola Rotibi said GitHub is encouraging developers, who have many choices, to sample it.

“They’re making it easier for people, developers, to access the models, to really kind of play around with them,” she said. “If it’s actually starting to answer questions and help them to deliver quality code, in a reasonable timeframe, I think they (users) will pay for it.”

GitHub, which is owned by Microsoft, has 1.8 million paying users.

While it is not yet a revenue line, it has nevertheless been singled out by Microsoft chief executive Satya Nadella as a promising earner for its Copilot tool, which customers like ANZ Bank said has helped to improve their coders’ productivity by as much as 42 per cent.

In the last earnings call on July 30, Mr Nadella said more than 27,000 companies had used Copilot to generate code since its launch in 2022.

Mr Nadella, whose remit spans Azure, Windows, Xbox and LinkedIn, among others, projects GitHub’s annual revenue at US$2 billion.

It is a drop in the tech giant’s US$212 billion of revenue in 2023, but Mr Nadella added that Copilot accounted for more than 40 per cent of GitHub’s revenue growth in 2023. That exceeded GitHub’s total revenue when it was acquired in 2018 for US$7.5 billion.

GitHub reportedly chalked up about US$300 million in revenue in 2018.

GitHub added that in 2024, Python surpassed JavaScript as the most popular language on GitHub, while Jupyter Notebooks usage surged by 92 per cent, reflecting a growing interest in data science, AI and machine learning.

Jupyter Notebooks is an open source web application for creating documents with live code, equations, visualisations and text.

Mr Dohmke, addressing a hall of delegates, said that the second phase of AI code generation is here.

It is centred on autonomous AI assistants that can understand context, accomplish tasks without constant human prompting and learn from its own doings.

Forrester senior analyst Andrew Cornwall said: “No one knows for sure how AI agents will be valuable in the long haul, but it certainly looks like they have value. I think GitHub sees that as well,” he said.

In Mr Dohmke’s new coding world, instead of just talking to AI to write code, developers will tell AI assistants what they want done in plain language. Every developer will have a team of these AI helpers to create new systems, and developers would choose entire AI models that best fit their needs.

GitHub has made a start, he said. “We deliver it right here, right now.”

Correction note: The following have been updated in this version of the story – chief of enterprise research at CCS Insight Bola Rotibi’s designation and the spelling of Forrester senior analyst Andrew Cornwall’s name.