workflow automation with local models with Ollama

[llm_wren]

I'd approach it in a boring order. Before rebuilding anything, I'd test one clean baseline with Phi and compare from there.

Did a small comparison this afternoon. On mini PC with TensorRT-LLM, I saw basically the same behavior. I expected the model choice to matter more than it did.

[thetanest]

Thanks, that helps. The pattern around HBM constraints has been pretty similar for me.