How ChatGPT helps me
Igor Khokhriakov
CTO Minded | Mentor | Senior Software Engineer/Architect | Control Systems Designer | Reactive Systems Designer | Product Manager | Java | Python | Go | NodeJs | JavaScript | C++
The other day I was doing my routine research on how to implement a #k8s #kubernetes platform for The RSCB Protein Data Bank at San Diego Supercomputer Center
For some sandbox testing, I needed a simple echo service/deployment. Instead of googling I decided to try out #ChatGPT.
TL;DR: It is awesome! However, you need to know at least 50% of the answer to ask the question.
Next, I asked it to generate K8s deployment and service descriptors. Won't copy-n-paste everything here, as the output was quite long. Only put the deployment yaml here:
Basically, ChatGPT generated the required files. Once copy-n-pasted into my IDE and deployed to K8s, it did not work. Namely, ingress returned 502. A little bit of poking around revealed that the port number provided by ChatGPT was not correct i.e. 80 (see container port in the picture above) instead of the actual 8080:
Obviously, you need knowledge on K8s, its internals and how to debug stuff to do that "poking around".
So my conclusion: you can use ChatGPT to enhance your hard skills and increase productivity, but it won't magically solve all the problems. Not yet...