Investigate and Demo how Langchain approaches building Agents

[harshsikka]

Context: we want to understand different approaches/methodologies to building "agents". Langchain builds LLM based agents, we want to understand their approach.

Output: Prepare an overview of the Langchain approach to building Agents, and build/demo a basic agent.

[harshsikka]

@bfaught3 Helen was going to work with you on this, I think we're still trying to figure out how to add her to GH. Worth reaching out to her to coordinate, perhaps over discord.

Edit: @helenlu66 added!

[helenlu66]

presentation on LangChain here

[helenlu66]

@bfaught3

• [x] investigate autonomy e.g. autonomous AutoGPT in LangChain
• [x] dataflow diagram of what happens between prompt input and response output

@helenlu66

• [x] identify LangChain unique capabilities
• [x] demo of LangChain unique capabilities

[harshsikka]

@helenlu66 I totally forgot to comment this a few days ago, but would be useful to capture even small tasks like this as their own issues

[bfaught3]

Presentation on autonomous AutoGPT in LangChain here Demo of AutoGPT in LangChain here

[harshsikka]

https://github.com/ManifoldRG/Manifold-KB/issues/19

Calling the above out again, would be good to understand each separate thrust in its own issue.

[helenlu66]

closing this for now since the to-do items are completed. We can re-open later if we decide to look more into LangChain