Considering that Llama's license dictate you should prefix your model name with "Llama" if you create, train, fine tune, or "otherwise improve" an AI model based on Llama, it might happen sooner than expected :) Fast-forward a couple of generations with their own similar license and maybe we'll end up with `LlamaRamaPastureForageGrazer/20280415`

I’m waiting for the model called “dwim-final”.

Hilariously, you can ask ChatGPT about OpenAI's model names, what they mean, and then have it suggest better ones. And they are.