This contributor uses local AI for home-lab experimentation and coding projects, with a particular focus on making low-end hardware more useful through customization. They build agents that use tools and skills, and use MCP tools to extend the workflow. Different Gemma variants run through LM Studio at roughly 7–20 tokens per second depending on the model. The setup is explicitly a hobby and learning project: the contributor has not spent money on LLM services or tools and does not feel that larger cloud models provide enough additional value for these experiments.
VERIFIABLE SOURCE
View submission form Submitted anonymously via Tally · reference Ar0B8NB