The hottest Substack posts of Gary’s Substack

And their main takeaways
4 HN points 16 May 23
  1. Experimenting with running powerful LLM models on CPU without GPUs.
  2. Llama.cpp allows real-time inference on CPU, providing a trade-off for larger models with more available RAM.
  3. Integration of Google Search API to retrieve search results and smoothly integrate them into the LLM responses.
0 implied HN points 14 May 23
  1. Gary Linscott has a Substack newsletter coming soon.
  2. The Substack link is garylinscott.substack.com.
  3. You can subscribe to Gary's Substack newsletter.