To wrap up this project, I wanted to write a post with my final thoughts.
Overall, this term I learned many valuable lessons. First, AI is great at generating starter code for a task for you. In many cases, you will have to edit this code to integrate it properly and guarantee functionality, but it will at least get you started. Second, AI is very capable of writing tests for your software, which can be very valuable and save lots of time. Finally, AI does a pretty good job of explaining a concept. I felt like I learned quite a bit about Next, just from AI explanations.
There were a couple of main drawbacks of AI that I encountered this term. First, AI is terrible at debugging; this was probably the most frustrating this all term. Whenever I told it that something didn’t work, it would either give me vague steps on how to troubleshoot (spoiler: these never worked), or it would just change its code slightly, which never resolved the problem. I had to rely on actual debugging principles to fix my issues. Second, AI can sometimes provide outdated code or information. Many times this term, ChatGPT generated Next.JS code that was so outdated, that Next threw specific error messages pointing out code that was written using out of date practices. Given how quickly technology changes, this can be an issue. Finally, AI is only as good as your query. You have to be very specific and very clear when asking the AI to do things, and there is a learning curve associated with this.
I learned so much this term, and I am happy to have worked on this project. Shoutout to my fellow development team members for making it a great experience!
Leave a Reply