The realm of coding is witnessing a transformative shift as AI becomes an increasingly integral part of the software development process.

GitHub.com – a Microsoft product, is a place that stores source code, in business terms, it can be compared to document management systems as it offers a platform for managing and collaborating on digital content. At present, it hosts a vast collection of over 100 million repositories. These repositories contain various types of content, such as code, projects, documentation, and related materials. Both enterprises and private individuals use it, in fact, GitHub is used by almost every developer (of the current 27 million developers globally). In recent years GitHub has been developing AI tools such as CoPilot – an AI coding assistant. Its aim is to write code significantly faster than a developer could.

Recent statistics reveal that GitHub Copilot is now responsible for a staggering 46% of all code written across programming languages on the GitHub platform. When it comes to Java programming, this number leaps to an impressive 61%.

In February 2023, GitHub Copilot saw a substantial upgrade to its underlying Codex model. This enhancement resulted in improved code synthesis, better context understanding, and a feather-light client-side model, marking a giant stride in its evolution. Many developers start out using Copilot for writing unit tests for their code, it’s almost like a gateway drug into AI for coders. This alone can reduce development time by at least 30% and even as high as 50%. Furthermore, Copilot now features an AI-based vulnerability prevention system, which in real-time identifies and blocks insecure coding patterns. Even if you don’t use it to write your code, having an AI watchdog on the lookout for vulnerabilities is undoubtedly worth its weight in gold.

Andrej Karpathy, a leading figure in AI development, formerly the Director of AI at Tesla, and the mastermind behind an ambitious project aimed at creating a JARVIS-like AI at OpenAI, has some valuable insights. In December 2022, he revealed:

“Copilot has dramatically accelerated my coding, it’s hard to imagine going back to manual coding.”

Andrej Karpathy

With AI now composing roughly 80% of his code with an accuracy of about 80%, Karpathy spends more time overseeing, prompting, and editing than coding from scratch. This was several months ago, and before CoPilot’s code was significantly improved.

This shift in the development process raises intriguing questions about the future of software engineering. With AI tools like Copilot and Amazon’s CodeWhisper becoming further integrated into the development process, software engineers’ roles are likely to undergo a significant transformation. As AI shoulders a larger share of the coding workload, engineers can focus more on refining, editing, and providing creative input.

In the near future, we may see a new breed of software engineers who excel in collaborating with AI systems to create efficient, innovative, and reliable software. These engineers can leverage the power of AI to streamline coding tasks, free up more time for creative problem-solving, and ensure the final product meets the evolving needs of an increasingly digital world.

As we navigate this future, embracing these changes and adapting to this new landscape will be key. For both current and aspiring developers, it’s about having the right mindset and a willingness to evolve. In doing so, the future of software engineering could be more efficient, creative, and dynamic than ever before.

If you’re a newcomer contemplating whether learning to code is worth the effort, the answer is layered with complexities. But as a senior technology expert for a large global enterprise, my answer leans toward, “Yes, it indeed is.” However, don’t forget to broaden your horizons. Mix in a dash of data science and business analysis into your learning journey – it could be highly beneficial.

What are your thoughts on the role of AI in software engineering? How do you see the industry evolving over the next five years? Let us know in the comments below!