You can't attack other users here and we've warned you many times. If you keep it up we're going to have to ban you.

https://news.ycombinator.com/newsguidelines.html

I didn’t think gp was recommending writing every thing from scratch, but rather advocating for more piecemeal tools that can be refactored or replaced, rather than buying into a whole framework. Those libraries you mentioned are much lighter weight. I like using guidance partially since it provides a unified interface to language models, but it’s not trying to do too other than managing prompts, output and iteration within prompts

Can you back that up?

I've looked at the documentation for both Microsoft Guidance and LMQL and they look like LangChain to me: frameworks where I'd have to spend a whole lot more time learning how to use them than if I just imported the OpenAI Python library directly and started running prompts through my own thin set of custom functions.

It's kind of strange to lump both of those (far simpler) "frameworks" (more like template standards) with LangChain. Indeed, they do seem to provide far more value than LangChain does.

Also:

> Gosh you are so fking wrong it's sad.

Isn't very civilized or proportional response.