“Brain Storm” JavaScript code
I have a subscription and use ChatGPT Plus to “brain Storm” JavaScript code for a mix of projects.
The code mainly runs using Node.JS.
I found that giving ChatGPT a very defined task the code results are very good.
I keep my requirements to small functions so that the resultant code content is less than 80 lines.
“Just Stops”
However, many times when the code starts to get near the end it “Just Stops” many times I can code it, but its quite a let down of such a smart system.
chatGPT retry bug
If you ask chatGPT to continue from the “for loop x” for example it will re-write the code again and hang in about the same place. So you have to teach it, Well give it some rules,
Work around and other bugs
My fix is to let it code till it stops and hangs and then type
“Your code is not finishing to the end” can you code in sections? … (GPT) Sure,
It will do the same bug and re-type the code again from the start and hang, (Showing its rule base not AI issue)
Type "No failed!, type in sections as code from the “for loop x” keeping the same context till the end,
Very important to add “Sections”, as “code” as GPT has another bug as it will forget to style as code and give you code as chat text.
The other issue is GPT needs to follow the algorithm its creating from the beginning to compute the full context. So typing “keeping the same context till the end” means that value X may not become Y.
Summery
I have read there is a line limit, but some times it hangs after just a few lines, Also I have paid chatPlus so its faster with less restriction. I’m still very impressed with my “Brain Storms” with chatGPT you just have to work with it.
I have found that when its made a coding mistake and you suggest a different approach it builds on it in ways you might have not used. I have been very impressed with using MySQL and Asyinc in Node.js which has some tricky code issues that it does better than us humans!
Also I know this sounds strange, I find my self writing, Good, great well done, just in case the AI team are going to use positive reinforcement in the future.