Back
Similar todos
🤖 played around with adding extra context in some local Ollama models. Trying to test out some real-world tasks I'm tired of doing. 
#research
🤖 Tried out Llama 3.3 and the latest Ollama client for what feels like flawless local tool calling. #research
🤖 played with Aider and it mostly working with Ollama + Llama 3.1 #research
🤖 spent my evening writing a better console for some more advanced Ollama 3.1 projects. #research
🤖 played with Ollama's tool calling with Llama 3.2 to create a calendar management agent demo #research
🤖 more working with Ollama and Llama 3.1 and working on a story writer as a good enough demo. #research
🤖 Created an Ollama + Llama 3.2 version of my job parser to compare to ChatGPT. It's not bad at all, but not as good as GPT4. 
#jobs
🤖 lots of AI research last night including writing a functional story bot to wrap my head around how to apply step-by-step logic and get something meaningful out of llama2 #research
🐍 🤖 used the Python llm and symbex libraries to write a bunch of code for me while my kids were taking a nap. Not bad. Not bad at all.
🤖 still sick but played with playwright and developed a pretty cool bot framework. might use that to automate more common tasks 
#mylife
🤖 dove a little deeper into a screenshots to code tool and the results are getting much better #research
🤖 got llama-cpp running locally 🐍
🤖 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
🤖 I realized I didn't need to chain multiple tool calls together for my Calendar Agent to work. I can have my create check and then stop overthinking it. #research
🤖 worked on old scraping project and making playwright act like a real person
🤖 testing Claude 3 for some low hanging fruit code writing tasks
🤖 worked on a better template/pattern library script tonight #mylife
🤖 built a new code generator tool when I wasn't driving today. #mylife
🤖 got sweep running on my Mac Studio and spent $0.73 letting it write five pull requests to try it out. Not bad.
📝 prototyped an llm-ollama plugin tonight. models list and it talks to the right places. prompts need more work.