Stop chasing parameter counts. Build the toolbelt instead. — What I learned building Tlamatini (Open Source Destktop App). (opens in new tab)

For the last few months I've been building Tlamatini, an open-source local-first AI developer assistant. Along the way I kept bumping into the same assumption — both in articles and in my own head — that to build something useful, you need the biggest model you can afford. GPT-4. Claude Opus. Llama 70B at minimum. Then I started actually shipping with smaller local models, and I learned something that flipped my thinking. The real lesson A 20B-parameter LLM, given the right tools, the right a...