does it run locally?
In theory it could -- but we currently still depend on modern LLMs due to their great latency vs locally run LLMs.