Skip to main content
David Loor
AboutServicesProjectsBlogContact
←Back to Blog

#Local Llm

3 posts with this tag

Found 3 posts
2026-06-18
4 min read

My agent kept losing facts it had already saved

AIdaemon could save a fact and then fail to find it when I asked for it. Semantic search ranks by what's on the same topic, so the short fact holding the real answer gets buried under wordier ones nearby. A reranker reads each candidate against the question and puts the right one back on top.

aisoftware-developmentopen-source
2026-06-08
13 min read

Local Gemma was too slow with AIdaemon until I fixed llama.cpp and the prompt size

I wanted AIdaemon on local Gemma 4 26B through llama.cpp, not Ollama. Generation ran at ~45 tok/s on an M4 Pro. Agent turns still felt stuck because prefill on 14k-token prompts took 8 to 9 seconds before the model wrote a single word.

aisoftware-development+1
2025-11-01
9 min read

How to run open source LLMs (AI) on your computer?

Learn how to run open source AI models (LLMs) directly on your computer. No cloud, no subscriptions, complete privacy. A beginner-friendly guide using llama.cpp.

AITutorial+1

Stay Updated

Get the latest posts and insights delivered to your inbox.

Unsubscribe anytime. No spam, ever.

Blog archive
  • My agent kept losing facts it had already saved
  • Local Gemma was too slow with AIdaemon until I fixed llama.cpp and the prompt size
  • How to run open source LLMs (AI) on your computer?
David Loor

AI, Cloud & Web Solutions Architect

AboutServicesProjectsBlogBookshelf

© 2026 David Loor. All rights reserved.

davo20019@gmail.com