So basically, you’re trying to use stuff that came later than the data set ChatGPT was trained on and complaining it doesn’t know things that are happening in the future..
But how will this affect what devs work on in the future? Will new frameworks be so hampered by this that no one will use them? I tried asking this a few days ago but got little response:
> If I ask some LLM/GPT a react question I get good responses. If I ask it about a framework released after the training data was obtained, it will either not know or hallucinate. Or if it's a lesser known framework the quality will be worse than for a known framework. Same with other things like hardware manuals not being trained on yet etc.
> As more and more devs rely on AI tools in their work flows, will emerging tech have a bigger hurdle than before to be adopted? Will we regress to the mean?
Which presumably will become a problem when gpt can use old scraper data or data full of gpt output and effectively starts eating itself and producing even worse results?
It's knowledge of sveltekit still seems weighted to pre 1.0, it hallucinates libraries for shadcn, or radix-ui.
You have to prefix your prompts with "Please use modern JS" otherwise it won't use .filter or .reduce sometimes.
I agree its a good replacement for Stackoverflow, but it is often pitched as a replacement for coding.
I've found using aider as the best compromise, let it build something small and revert if it doesn't work.