Feels like there should be value in building smaller, more specialized models - maybe even doing so on-demand. I don’t always want a model that knows Polish and

Feels like there should be value in building smaller, more specialized models - maybe even doing so on-demand. I don’t always want a model that knows Polish and astrophysics and Shakespeare, I want one that runs really fast and is laser-focused on the domain that I’m working on. I want to be able to say to a large general purpose LLM: “write a script that trains a model that is optimized for <useful task>” and then run that model. Edit: well gosh darn. Within the edit window for this comment, Google goes and launches Gemma 3 270M.