SmallTalk LLM
What is SmallTalk LLM?
Research into the smallest language model for natural conversation.
SmallTalk LLM is Anson Washeck's research project exploring how small a language model can be while still supporting natural, casual conversation. It sits alongside his broader work in local-first AI and practical small-model systems.
- Focuses on compact language-model design for natural conversational interaction.
- Explores the tradeoffs between model size, useful dialogue behavior, latency, and local deployment.
- Connects to Anson's Edge AI practice: capable models that can run closer to the user and within tighter hardware constraints.
24.1M-parameter local semantic model · WebGPU/WASM · no answer server