Hacking LangChain for fun and profit
blog.kevinhu.me
Hacking LangChain for fun and profit
1–10 of 27 posts
Re: Hacking LangChain for fun and profit
#2Great conceptual explanation of the different primitives in LangChain. Some possible inspiration for your upcoming piece involving vector stores with LangChain - https://milvus.io/blog/conversational-memory-in-langchain.md
Re: Hacking LangChain for fun and profit
#3still not sure of the value of the langchain abstraction. does anyone use it in prod?
Re: Hacking LangChain for fun and profit
#4still not sure of the value of the langchain abstraction. does anyone use it in prod?
Some of the abstractions seem a bit forced, but it seems nice to be abstracted away from how you'd handle vector storage or the LLM that get used (maybe someone launches something that does better than GPT4 on some sets of tasks and a few line change to swap over would be nice).
Re: Hacking LangChain for fun and profit
#5There's a pretty steep learning curve to langchain. I recommend avoiding it if you don't need higher level of abstractions.
Re: Hacking LangChain for fun and profit
#6still not sure of the value of the langchain abstraction. does anyone use it in prod?
We do, but mostly as a fairly nice and fairly well-documented interface for plugging together tools like vector DBs and memory with LLMs.
My feeling is that with the agentic/AutoGPT hype fading and OpenAI adding the functions API there's a lot less value in LangChain's abstractions, at least for production use cases. They're still cool for hacking/toys.
Re: Hacking LangChain for fun and profit
#7I’ll save yall some time. If you are familiar with concatenating strings in python, langchain is a library which does just that, only a little less cleanly. Hope this helps!
Re: Hacking LangChain for fun and profit
#8still not sure of the value of the langchain abstraction. does anyone use it in prod?
We do, but mostly as a fairly nice and fairly well-documented interface for plugging together tools like vector DBs and memory with LLMs. My feeling is that with the agentic/AutoGPT hype fading and OpenAI adding the functions API there's a lot less value in LangChain's abstractions, at least for production use cases. They're still cool for hacking/toys.
A little premature don't you think? Langchain is only a few months old
Re: Hacking LangChain for fun and profit
#9Earlier quoted context omitted.
We do, but mostly as a fairly nice and fairly well-documented interface for plugging together tools like vector DBs and memory with LLMs. My feeling is that with the agentic/AutoGPT hype fading and OpenAI adding the functions API there's a lot less value in LangChain's abstractions, at least for production use cases. They're still cool for hacking/toys.
A little premature don't you think? Langchain is only a few months old
It’s no longer useful / never was that useful. It’s main function was a shiny toy for investors.
Re: Hacking LangChain for fun and profit
#10I’ll save yall some time. If you are familiar with concatenating strings in python, langchain is a library which does just that, only a little less cleanly. Hope this helps!
That's a bit cynical. It does many other thing as well. For example, it exposes simple re-usable modules, also known as "functions" in Python. Sometimes it also uses Python "lists" by wrapping them into... classes.