This blog post is going to be a bit unusual, but stick with it and you’ll see why by the end.
A year ago, I wrote a blog (Low code, no code to AI programmer) about my experience using OpenAI, GitHub Copilot, and CodeT5, and how I thought these technologies would transform the lives of developers. In particular, how GitHub Copilot had made my development work more efficient.
Recently, there has been a lot of excitement and buzz around ChatGPT, a new way to interact with the OpenAI language model. ChatGPT is a powerful tool that uses advanced machine learning algorithms to generate human-like responses to text inputs. Many people have been impressed by its ability to understand and respond to complex questions and conversational prompts, making it a valuable tool for a variety of applications.
In my opinion, ChatGPT makes it easy for anyone to use AI-powered tools, regardless of their technical expertise. By providing a user-friendly interface and clear explanations of how the technology works, ChatGPT democratizes access to these powerful tools and makes them accessible to a wider audience. This is a great example of how OpenAI is helping to make cutting-edge technology more accessible and understandable for everyone.
I took the opportunity to explore OpenAI capabilities further and look at solutions for making developer’s life more efficient. I built a simple VSCode extension (I called it Kanzen. Lookup the meaning in Japanese and decide for yourself if it is apt) that can perform a range of operations using OpenAI APIs. I wanted to write a blog to share my experience about it.
The VSCode extension calls OpenAI APIs and can perform any of the following which most developers don’t enjoy doing. With this extension, developers can easily access the power of OpenAI’s cutting-edge technology to automate these tasks and save time and effort.
- Create a test case for a piece of code that the developer has already written
- Add documentation comments to their code
- Refactor code
- Custom — perform any other custom request from the developer. For example write a MongoDB query given a sample document.
My VSCode extension may not be particularly unique or fancy, but it demonstrates how OpenAI is making sophisticated technology accessible to everyone. By providing easy-to-use APIs and tools, OpenAI is helping developers of all skill levels to incorporate state-of-the-art machine learning into their projects and create innovative new solutions.
You can clone it and take it for a spin. Kanzen GitHub . Let me know what you think. You need your own OpenAI API key!
Here are some of the results of my experiment with this tool.
# Automate creation of Test Cases
# Automate documentation
# Refactor code following functional programming style
# Given sample document, write MongoDB Aggregation pipeline automatically
The best thing is that I used ChatGPT to co-write this blog as well 🙂
While this is all exciting, you need to be aware of a few limitations.
- The most important of them all is that the responses are based on information fed during training process, mostly publicly available data — and these are not necessarily absolute truth or facts.
- ChatGPT sometimes writes plausible-sounding but incorrect or nonsensical answers.
- ChatGPT is sensitive to tweaks to the input phrasing or attempting the same prompt multiple times.
While there are several limitations of this technology in its current iteration, some of them listed above, I am excited about the potential of these technologies to revolutionize the way we work and create. As a developer, I have seen firsthand how they can make our lives easier and more productive. And as AI continues to advance, I am sure that we will see even more amazing innovations in the coming years.
