Comment by lithocarpus

I don't get the importance of the distinction really. Don't LLMs and Large non-language Models fundamentally work kind of similarly underneath? And use similar kinds of hardware?

But I know very little about this.