More talk about ChatGPT
Over the last few months there were tons of ChatGPT related posts showing up in my timeline, mostly on how you can use it to accomplish some task. Write a mail, generate code or do some research. Many valid use-cases and definitely a lot of potential, but also a lot of repetition. I usually found approaches more interesting where the API was used to create a new service.
Therefore I also wanted to tinker a bit with the API and see how you can work with it. And as I just had some fun with implementing an Apache Camel component for Zeebe, I just started on doing the same for OpenAI. I started with a little distraction on how to talk to the API in Java until I figured out the implementing my own HTTP client is the fastest and most productive approach, especially when you can ask ChatGPT to generate some working examples for you.
It is still work in progress and still needs mord tests and documentation but otherwise it is making good progress to start a pull request rather sooner than later and should make it easy to leverage OpenAI capabilities in camel routes, e.g. for moderation.
My learning: