Hacker News new | past | comments | ask | show | jobs | submit login

A good trick here is to use the API rather than the web interface and give it access to functions that let it edit files directly. Then you don't have to do the iterations where you ask it for something new and it forgets something old, it just updates the code in place, leaving all the old stuff and changing/adding the new stuff.



Interesting. Is there an IDE (ideally, VSCode) extension you could recommend for that?

What do you mean by giving access? Is it possible to limit its access only to certain functions/files?


Not the GP, but I've been working on an open platform [0] for integrating OpenAI APIs with other tools (VSCode, JupyterLab, bash, chrome, etc) that you might find interesting, the VSCode integration supports editing specific files / sections etc.

Also worth taking a look at Github Copilot Chat[1], it's a bit limited but in certain cases it works well for editing specific parts of files.

[0] https://github.com/lightrail-ai/lightrail

[1] https://marketplace.visualstudio.com/items?itemName=GitHub.c...


This is exactly the use that Promptr is intended for https://github.com/ferrislucas/promptr

* full disclosure: I’m the author of Promptr


Looks promising, thank you! I will try it out this week.


Yes. Cursor at cursor.sh. I'm happily paying for it, and it works great giving answers based on your codebase. It generates great inline code, but doesn't have file and multi-file generation (yet?).


Not built in, you have to use the API and then build functions for it to invoke which would fetch the files. It’s called function calling (https://openai.com/blog/function-calling-and-other-api-updat...) but it’s not as easy as you might expect.


cursor.sh. Does exactly that.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: