Why learn LangChain?

DanielOctober 18, 2024October 18, 2024

Learning new tech as a software engineer is like investing in a stock portfolio. Some will fail, but others will make most of your income.

This is one of my key takeaways from The Pragmatic Programmer book.

My current investment is in LangChain, and in this article, I want to explain why I believe LangChain is a worthwhile investment.

What is LangChain?

LangChain is a framework designed to make it easy for developers to integrate AI features into their applications. Just look at the NPM usage graph for LangChain.js, and you’ll see it’s rising rapidly:

And there’s a good reason for this growth!

Challenges with Large Language Models (LLMs)

Large Language Models (LLMs) such as ChatGPT, Llama, or Claude AI are complex tools for web developers. They operate in a random, unstructured, and nondeterministic manner. Here are some of the challenges we face when working with LLMs:

• They can give different answers or formats despite clear instructions.
• They can be slow and expensive.
• They are stateless and lack both short-term and long-term memory.
• They can't access external information beyond their training data.
• They can hallucinate and provide misleading information.

LLMs are extremely powerful, but to work with them efficiently, you need to understand their nuances.

How LangChain Helps

This is where LangChain comes in. The framework encapsulates all the abstractions and concepts you need when adding AI-powered features to a web app in one package.

Both Andrej Karpathy (ex-OpenAI and ex-director of AI at Tesla) and Swyx Shawn Wang (host of Latent Space, the top AI Engineering podcast) have mentioned that soon, there will be more AI Integration Engineers than traditional Machine Learning Engineers (source 1 and source 2).

LangChain has both Python and JavaScript implementations, which is great news for web developers. It allows us to leverage our existing JavaScript skills to communicate with AI models.

Why LangChain is a Solid Investment

The AI field is constantly evolving, with new models, vector databases, and embeddings emerging frequently. LangChain offers the flexibility to easily swap components in and out, making it an ideal foundational framework.

Companies like OpenAI and Anthropic are motivated to grow the number of AI Integration Engineers since they are the primary users and customers of AI models. Frameworks like LangChain have become the standard tools for AI engineering.

Can you work with AI models without LangChain? Absolutely. But using a well designed framework like LangChain will save you time, reduce frustration, and lower costs in the long run.

LangChain's Ecosystem and Community Support

I believe that LangChain is here to stay. Let's take a look at the numbers for the LangChain ecosystem:

• 📦 15 million monthly downloads
• 💻 100k apps powered by LangChain
• 🧑‍💻 2,000 contributors
• ⭐️ 90k GitHub stars
• 🤖 80 LLM providers supported
• 💿 100 vector store providers
• 🗂 100 document loaders

LangChain is backed by a solid company and a dedicated team. Multiple times I've had folk sfrom the core LangChain team reaching out to suggest some improvents to my articles or to help out. In very few projects I've seen this level of involvement.

It’s not a small GitHub project that might randomly stop being maintained. These numbers and dedication are my reasons to think that LangChain will be constantly updated and continue to evolve.

Conclusion

And there you have it, folks. LangChain is more than just a buzzword. With its growing ecosystem, strong community, and ability to adapt to the ever-changing AI landscape, LangChain offers the tools for a strong foundation.

In my opionion, it's clear that LangChain is here to stay, and it's a worthwhile investment.