At the end of June, GitHub unveiled a technical preview of a new tool called GitHub Copilot, an AI pair programmer that lives inside the Visual Studio (VS) Code editor and autocompletes code snippets. That’s right—the use of AI in coding is here.
The idea behind pair programming is that two coders work on the same project to catch each others’ mistakes and speed up the development process. With GitHub Copilot, one of those coders is virtual.
The system draws on source code uploaded to the code-sharing service GitHub, which was acquired by Microsoft in 2018. Microsoft and GitHub developed the product with the help of OpenAI, an AI research startup co-founded by Elon Musk.
For now, the product is still in the limited technical preview stage. Its first implementation inside a VS Code extension is being tested by a small group of users. If this technical preview doesn’t reveal any major wrinkles, GitHub plans for it to become publicly available for VS Code and Visual Studio as a commercial product.
Here’s a quick look into some of the things GitHub Copilot can do:
- Analyze the code already written and generate new matching code, including specific functions that were previously called.
- Assemble code for a user when the user describes the logic they want in a comment.
- Learn to match a user’s coding style based on the user accepting and rejecting the AI tool’s suggestions.
- Autofill repetitive code.
- Suggest tests that match a coder’s implementation code.
Is AI replacing us?
Of course, this exciting new development begs the question: If these AI tools are doing our work for us, how long will it take for them to do our work in place of us?
AI is developing faster than ever and is expected to snowball in the next few years, reaching a market value of $190 billion by 2025. For reference, the market value was $62 billion in 2020.
New AI tools come out regularly, including content writing tools that use artificial intelligence to conduct research about a given keyword, read articles about it, and then write SEO-friendly, plagiarism-free content. A few tools that use this technology include Writesonic, Article Forge, and WordAI.
And while on the surface it appears human writers (and coders) seem to be in danger of being rendered useless, that isn’t necessarily the case. When it comes to writing, there are still things AI can’t do. For example, it can’t create truly original content or show creativity, empathy, or sound judgment.
An AI-powered coding tool is nowhere close to replacing a human coder. It has no way of determining which features to prioritize or what problem a piece of software in development would address.
In an FAQ about the service, GitHub said, “GitHub Copilot doesn’t actually test the code it suggests, so the code may not even compile or run … GitHub Copilot may suggest old or deprecated uses of libraries and languages. You can use the code anywhere, but you do so at your own risk.”
Yes, Copilot will get better with further development, but still, yikes.
Friedman himself weighed in on the debate of AI tools replacing humans, stating:
“We think that software development is entering its third wave of productivity change. The first was the creation of tools like compilers, debuggers, garbage collectors, and languages that made developers more productive. The second was open source where a global community of developers came together to build on each other’s work. The third revolution will be the use of AI in coding. The problems we spend our days solving may change. But there will always be problems for humans to solve.”
So for now, we can rest easy about our jobs getting swept out from under us and enjoy the conveniences AI tools have provided.