Learn from LangChain, BUT Don’t Build With It!!!

preview_player
Показать описание
There is a lot of buzz around new large-language model frameworks like Lanchain. Unfortunately these tools are too immature for production use because the underlying models and best practices are changing too quickly.

Instead of building with these frameworks, review their open source code for inspiration. Often they’ve developed clever ways for interfacing with LLMs. And also read the content and case studies these teams produce.

They are at the cutting edge and sharing valuable insights. Our favorite is LlamaIndex. LLM frameworks are the future, but they probably won’t be ready for a year or two. In the meantime, do what other teams are doing: build your own interfaces to LLMs.

For more check out our channel @prolegoinc
Рекомендации по теме
Комментарии
Автор

already a year since this video how is it now?

LuisJavierStudio
Автор

in other words, ship in 2 years instead of 2 weeks.

hmaina
Автор

Im learning, but i want to start simple, though, and understand the code as much as possible. I wish there was a good rag tutorial with full code explanatipn.

Larimuss
Автор

What are some interesting interfaces for LLMs you have seen recently?

vandaloart
Автор

Please could you make a video explaining how is the current artificial intelligence engineering today? 5 months ago you had to create your own model with pytorch, create your own dataset, train the model in google colab, test it and test it until the model worked well, 4 months ago you didn't have to do that, there were already pre-trained models in Huggin Face, you just had to fine tune them with your own dataset, then came out RAG and a wave of great LLM models trained and with easy access with APIs, that is to say with an api to gpt4 and sending your data as embedings/tokens now you can do much easier what 5 months ago was done with much more work, so I would like to know how engineering in artificial intelligence works today.

RingOfFaia
Автор

i am still writing in Go, and just using raw prompt creation.

robfielding
Автор

What about haystack? What framework should I use to build an application for companys

orlandocastellanos
Автор

Checkout the Eliza Framework y'all!

ccerrato