created several personal AI agents via ollama #leifinlavida

implemented ollama local models #rabbitholes

🤖 Updated some scripts to use Ollama's latest structured output with Llama 3.3 (latest) and fell back to Llama 3.2. I drop from >1 minute with 3.3 down to 2 to 7 seconds per request with 3.2. I can't see a difference in the results. For small projects 3.2 is the better path. #research

🤖 more AI writing project

✏️ I wrote and published my notes on using the Ollama service micro.webology.dev/2024/06/11…

🤖 worked on a random automation project with copier #mylife

✏️ wrote about running Llama 3.1 locally through Ollama on my Mac Studio. micro.webology.dev/2024/07/24…

🤖 dove a little deeper into a screenshots to code tool and the results are getting much better #research

🤖 worked on a better template/pattern library script tonight #mylife

I used Claude 3.5 Project + Artifacts to help refactor most of the Ollama + Llama 3.1 #research project. I would call it a writing bot, but I'm not building it to automate writing. It's mostly a content wrapper around the Chat interface but it's good at generating code and building off of chat history.

prototype a simple autocomplete using local llama2 via Ollama #aiplay

🤖 more robot work, but figured out how to get better results limiting num_predict for git commit messages #research

🤖 built a new code generator tool when I wasn't driving today. #mylife

🤖 has Claude work on a few projects while I was going to sleep and immediately after waking up. The present is wild. #research

Experiment with AI 🧪🧠

🤖 tried out ChatGPT to mixed results

talk about LLM finetuning and Alpaca and ChatGPT being just instruct-text model #life fxtwitter.com/levelsio/status…

🤖 sat down and had ChatGPT write some GPT4(ish) code for me

🤖 wrapped my brain around proxies and built an app which collects and tests them #mylife