🔄 2003: parsing garbage HTML. 2026: parsing garbage LLM output. Full circle. - https://swm.cc/notes/the-llm-is-the-new-parser
🤖 Local LLMs can see family photos.It even detected the beer brand. https://swm.cc/writing/indexatron-local-llm-photo-analysis/
swm.cc
Indexatron: Teaching Local LLMs to See Family Photos
An experiment using Ollama with LLaVA and nomic-embed-text to analyse family photos locally - proving that privacy-preserving AI photo analysis actually works
Lil science experiment this Sunday morning on AI Photo analysis API using models via ollama... 👨🔬🧪🥼
Wee Saturday night hacking session on a few side thingies is always a good shout.. Good way to spend <3 day
'lil blog on what I did this avo... Feels good to do this now and again.. https://swm.cc/writing/building-swanson-recommendation-system/
swm.cc
Building Swanson: A RAG-Enhanced Recommendation System for whatisonthe.tv
How I built a context-aware recommendation system using Claude, Server-Sent Events, and viewing history to suggest TV shows and films
Proving out to be a nice wee addition to this project...
Doing a bit of AI programming now - well just a wrapper to an AI agent to help me do things will document if a success. Time will tell.
The best AI tools are the ones that get out of your way. Give me results, not a conversation.
