It kind of doesn't matter. Llama is great because it lets everyone experiment, optimize and learn with LLMs locally. Once Stability releases their model everyone will just drop llama and quickly apply all their new tools and learnings to that model.
There is a risk of the community getting "entrenched" in an old model/architecture.
In stable diffusion land, for instance, the community is pretty much stuck on the "old" architecture. Newer innovations, like Huggingface diffusers and various optimizations derived from that like PEFT, torch.compile support, AITemplate/TensorRT compilation and various other bits are largely unused.
They are also pretty much stuck on SD 1.5, even though 2.1 is a good base for finetuning.
