GitHub: AI for developers boosts global productivity
- AI for developers productivity could boost global GDP by over US$1.5 trillion by 2030.
- GitHub anticipates that approximately 80% of code will be written with AI.
- Some developers are concerned about where the code is coming from.
AI for developers is reportedly boosting productivity levels much more than initially expected. Today, there are more than a million developers using AI to generate code. In fact, GitHub Copilot has been adopted by over 20,000 organizations with more than three billion lines of code developed.
GitHub Copilot is an AI pair programmer that offers autocomplete-style suggestions as developers code. Developers can receive suggestions from GitHub Copilot either by starting to write the code they want to use or by writing a natural language comment describing what they want the code to do. GitHub Copilot analyzes the context in the file they are editing, as well as related files, and offers suggestions from within the text editor. GitHub Copilot is powered by OpenAI Codex, a new AI system created by OpenAI.
With the release of OpenAI’s GPT-4, GitHub Copilot X was launched, providing evolved chat and voice interfaces as well as support pull requests, answer questions on docs and offer a more personalized developer experience.
As the world’s most widely adopted AI tool for developers, GitHub has released new research detailing the potential economic impact and productivity benefits of generative AI. The study found that AI developer productivity benefits could boost global GDP by over US$1.5 trillion by 2030.
Here’s a summary of the findings from the Sea Change in Software Development report:
- Less than a year after its general availability, GitHub Copilot is turbocharging developers writing software. Analysis of a large sample of GitHub Copilot users (n = 934,533) reveals a sizable productivity impact. On average, users accept nearly 30% of code suggestions and report increased productivity from these acceptances. Furthermore, this productivity impact increases with time, and the benefits are greatest for less experienced users.
- The research estimates that these generative AI developer productivity benefits could boost global GDP by over US$1.5 trillion by 2030 by helping to meet the growing demand for software. GitHub acknowledges that these estimates are conservative as they are moment-in-time projections that do not account for the increased demand for software development due to its greater efficiency and continued digital transformation that will arise from generative AI adoption.
- Another interesting finding is that the global landscape of players working on generative AI is diverse, including big tech, start-ups, academia, and individuals. Open source activity on generative AI has seen an exponential increase compared to previous years, based on an analysis of GitHub repositories and commits. Findings suggest that the open-source ecosystem, particularly in the United States, is driving generative AI software innovation. Individual developers are leading the majority of such repositories on GitHub.
According to Thomas Dohmke, GitHub CEO, the economic impact of generative AI over the next decade will be profound and the world is already seeing large-scale adoption of AI coding tools like GitHub Copilot by developers and companies. In a recent survey, 92% of developers said they use AI tools both in and outside of work, which underscores how quickly these tools are redefining the overall developer experience.
“What we draw from all this is that generative AI is turbocharging developer productivity with gains that will ultimately drive a boom in GDP for the global economy and, in turn, a surge in demand for software developers. We’ve seen this throughout the history of developer tool innovations from compilers to open source, and we’re already seeing that again with GitHub Copilot and soon GitHub Copilot X. One year later, we’ve realized this collision of AI and the software developer will not lead to a decrease in developer jobs — it will lead to AI augmenting developer potential and accelerating human progress,” commented Dohmke.
Less experienced developers benefit more from GitHub Copilot
GitHub’s study found that less experienced developers have a greater advantage with tools like GitHub Copilot, which is corroborated by other studies, including GitHub’s previous experiments on the impact of AI on developer productivity. As developers use these tools to upskill, they will become more fluent in prompting and interacting with AI to power the development lifecycle.
As a matter of fact, by examining GitHub Copilot telemetry, the study found that users accept an average of 30% of code suggestions, representing real productivity gains. Over time, developers overcome the learning curves and get comfortable with the tool. This leads them to use it for an even greater impact and accept more code suggestions. They are also consistent with and extend previous experiments and related productivity research on generative AI.
“Accepting code completion suggestions is beneficial for developers as it allows them to finish writing a code block faster and can save time searching for less-commonly used syntaxes. Acceptance rate captures code immediately adopted from GitHub Copilot suggestions based on the active choice of the developer. This is confirmed in survey research, which finds that developers’ statements of productivity are closely tied to acceptance rate,” the report stated.
In terms of productivity gains, the data from the study indicates that AI for developers is most likely going to be the way forward. As developers continue to become fluent in prompting and interacting with AI, particularly with new models that allow natural language to power the development lifecycle, GitHub anticipates that approximately 80% of code will be written with AI.
Put simply, GitHub believes that this will help democratize software development, allowing more people from non-technical backgrounds to build and shape the software that will continue to power the global economy.
Where is the code for AI for developers being developed from?
Here is where it does get a bit interesting. As GitHub clearly indicates that AI for developers is becoming increasingly common, there are some questions as to the legitimacy of the codes being developed.
According to several blogs and posts in forums, many developers feel that while GitHub’s Copilot is making things easier, what it is actually doing is just scanning different projects on it and giving developers code that already exists. Put simply, some developers feel that the AI is not really learning how to code or create new code but instead, it just steals existing code.
And what makes it more annoying is the owner of GitHub, which is Microsoft. As a profit-driven company, developers feel the code should be generated and not taken from other projects. At the same time, as the code is from Copilot, they would need to have the right to use the code.
While there are also arguments that Copilot is open-source, when it comes to AI for developers, the understanding is that attribution needs to be given. Tech Wire Asia also uncovered several blogs that are discussing similar issues. While this is not going to slow the use of AI for developers, some would still be reluctant to fully use the code unless they can see its source.
The end goal
GitHub’s study also acknowledges that the economic value of open-source activity will continue to grow, and will only be further accelerated by AI. But this would mean a growing need for both developers. Currently, there are only an estimated 27 million professional software developers in the world. But these figures are far from the amount of developers actually needed.
In Europe, the shortage of IT jobs is estimated at over 500,000, with Germany exceeding 100,000, and countries like Denmark, Norway, and Sweden are expected to face growing demand. Latin America faces a similar shortage as well, with a deficit of IT workers in Brazil growing to an estimated 25,000 per year. An estimate by Korn Ferry states that the shortage of tech talent could be 85.2 million in 2030 across various sectors.
GitHub believes this is where Generative AI tools can make a difference. Not only can it be used to skill more developers, but also to make each developer more productive. It thus holds promise to meet the exploding demand for software that will only accelerate as AI adoption continues.
“From clear gains in productivity to improved acceptance rates as time goes on, to the explosion of generative AI in open source repositories, to the potential US$ 1.5 trillion to the global GDP, these findings signify a sea change in software development. As more developers adopt these tools and become fluent in the skill set of prompting with generative AI, it is clear that this new way of software development has created an inextricable link between humankind and artificial intelligence that could well define how the world’s software is built for generations to come,” concluded the study.
- The end of TikTok Shop and other social commerce in Indonesia
- Lost in translation: Can AI tools improve?
- Is ChatGPT enabling collaborative decision-making or merely Hobson’s choice?
- NVIDIA and NTT DOCOMO revolutionize telecom services with world’s first GPU-accelerated 5G network
- Sony battles new hack: ‘Is my account safe?’ Echoes among concerned customers