How ChatGPT Plugins Might Create a More Collaborative Environment
Redefining teamwork
AI Disclosure: This article was co-written with ChatGPT. Since I am not a native English speaker, I have used artificial intelligence (AI) techniques to enhance the quality of the language used, the readability of the text, and the organization of the piece as a whole.
Welcome to a world where artificial intelligence is transforming the landscape of software development, and collaboration is the name of the game. ChatGPT can already do things that amaze even the most experienced coders daily (some naysayers aside). But now, plugins are entering the stage. Think of ChatGPT as the iPhone and plugins as the App Store. This would create an ecosystem full of creative possibilities.
As this article will raise some eyebrows (especially from people in the enterprise world, wait for the end): I will also look at the enterprise setting and how the things said in this article will still apply there to some degree.
The Appearance of ChatGPT Plugins
OpenAI has just released the limited alpha version of ChatGPT plugins. These innovative add-ons connect ChatGPT to third-party apps and let developers create APIs that make ChatGPT even better than it already is. By using these APIs, ChatGPT can now do a lot of different things, like get real-time information, access knowledge bases, and take actions on behalf of users. Even though the ChatGPT plugin system is still young, it will likely change a lot as OpenAI works with users and developers to improve and grow the plugin ecosystem.
ChatGPT Plugins Make AI Development More Open
The introduction of ChatGPT plugins marks the beginning of a new era of open AI development. This is similar to the early days of the App Store, when even small projects could do very well. With ChatGPT plugins, developers with different skill levels can now add to the AI ecosystem by making custom features that make ChatGPT more useful.
This paradigm shift lowers the barrier to entry for AI development, granting developers the opportunity to experiment and innovate without the need for extensive AI expertise. By fostering a development community that is diverse and open to everyone, we can use the power of different points of view to solve hard problems and come up with innovative solutions.
How ChatGPT plugins affect the software development process
Even though ChatGPT plugins have the clear potential to change how software is developed, it’s important to give specific examples to grasp their impact. Plugins that create test cases or analyze algorithms, for example, can boost productivity and make workflows much easier, and they are only a small step away. The models themselves are already capable of it, and you can generate test cases for almost anything directly via GPT class models or via tools like Tabnine. This step is more about the process and the user interface than it is about really improving the AI level. And this is exactly where plugins can help out.
The next interesting point is that we will see plugins that make the general “back and forth” with the AI easier. Usually, as a human, you are in charge of overseeing the process and coordinating the results. You also have to fix small mistakes that the AI made. Here, we also need better user experiences and improved workflows.
Collaboration plugins
However, one key area for growth is the world of developer collaboration plugins, which make it easy for multiple developers and their AI counterparts to work together in the same environment. By making AI an important part of the software engineering team, developers can use its skills to create functional components, classes, services, and patterns, similar to how 3D printers give hardware engineers the physical building blocks they need.
I don’t know exactly (yet) how these plugins will evolve to look. But I think Midjourney has done more in that direction, as the public recognizes now. For whatever reason, they selected Discord from the very beginning as their primary way to act with the model. And I don’t know if you've played around with that yet, but it’s super awesome to see the artwork from other people, happening in real time, their process, their interaction with the model.
Sometimes, if I have a good prompt, somebody will use part of my prompt in their prompt. Sometimes the reverse happens. After all, everything is public.
With these collaborative plugins, developers can work with AI models to improve and optimize the code that the AI generates to fit their needs. So, developers can focus on high-level design and decision-making, which lets them make more innovative and stable software solutions. But it’s important to be aware that AI-generated code may need a lot of human help to reach the quality and functionality you want.
And this, again, is getting easier thanks to the real-time collaboration of multiple people.
Also, adding ChatGPT plugins to the software development process helps developers be more creative and come up with new ideas because they can try out different AI-generated components and design patterns. If your overall architecture is good (!), it will be super easy to mock/test new stuff by adding strategies to your strategy pattern or letting the AI generate a new object for an already existing factory in your code (these “you have X, now do Y accordingly" tasks that right now are typical learning tasks for junior developers can just be automated. But don’t worry, we will find new tasks for our young colleagues ;-).
Wrapping Up — And Looking at the Enterprise Setting
In short, I think ChatGPT plugins could have a significantly positive effect on how software is made (but the world is changing so fast these days, who really knows in the end?).
I think, at least on a conceptual level, the impact of these ChatGPT plugins on software development processes cannot be overstated. Of course, in a big enterprise setting, you will never run your source code through OpenAI's AI; you will instead use your own self-hosted system. At some point, this will be as normal as setting up your own Kubernetes cluster.
Nonetheless, the concept of embedding “apps” into your self-hosted LLM interface would be something you most likely still want to have. Here, we can learn from these early takes, and the open-source community can, with our own help, catch up over time and build the foundation for future enterprise settings.
