tlowe11

4 ms·

1. 1.▲
   by tlowe11 3y ago

   What's covered: What’s memory for LLMs and why is it important? Are larger context windows always better? 3 techniques to manage memory. Using memory with retrieval systems.
2. 2.▲
   by tlowe11 3y ago

   Motorhead doesn't use Langchain under the hood, but you can use Motorhead with Langchain if needed – check out the docs here https://python.langchain.com/en/latest/modules/memory/exampl...
3. 3.▲
   Show HN: Hosted Motorhead (LLM Memory Server)(getmetal.io)

   6 points by tlowe11 3y ago|3 comments
4. 4.▲
   Show HN: Motörhead, LLM Memory Server Built in Rust(github.com)

   16 points by tlowe11 3y ago|1 comments
5. 5.▲
   by tlowe11 4y ago

   Great to hear! Humbled by the Firebase comparison as well :) We've talked about the markdown/HTML feature, definitely something we want to build. And I'll ping you!
6. 6.▲
   by tlowe11 4y ago

   Fantastic question. It's a few years old but I still love Infest the Rats Nest by King Gizzard. Especially awesome considering they're not a full time metal band. I'll give Periphery V a spin tonight!! Thanks for sharing :D
7. 7.▲
   by tlowe11 4y ago

   Thank you! We have an OCR pipeline already so you can upload the files and we’ll process them, chunk the text, create the embeddings and index them. Right now, we support PDFs, but the pipeline is ready to accept images as well. We’re openi
8. 8.▲
   Launch HN: Metal (YC W23) – Embeddings as a Service

   196 points by tlowe11 4y ago|99 comments
9. 9.▲
   by tlowe11 4y ago

   Our company's most recent round of raises were below inflation on the year