Evidence receipt / recommendation
Published · transcript-backedShawn Wang: recommendation
6 Sept 2023 Latent Space The Point of LangChain — with Harrison Chase of LangChain
“I have said that that piece of research is the best bull case for Lang chain and all the vector companies, because it means you should do chains.”
Source trail
Everything needed to verify it.
- Speaker
- Shawn Wang
- Attribution
- Verified speaker
- Claim type
- recommendation
- Recorded
- 6 Sept 2023
- Publisher
- Latent Space
Transcript context
…I think it's real. Basically I've found long context windows really good for when I want to extract like a single piece of information about something basically. But if I want to do reasoning over perhaps multiple pieces of information that are somewhere in like the retrieved documents, I found it not to be that great. Yeah. I have said that that piece of research is the best bull case for Lang chain and all the vector companies, because it means you should do chains. It means you should do retrieval instead of long context, right? People are trying to extend long context to like 100K, 1 million tokens, 5 million tokens. It doesn't matter. You're going to forget. You can't trust it. I expect that it will probably get better over time as everything in this field. But I do also think there'll always be a need for kind of like vector stores and retrieval in some fashions.…
Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.