Q: How open?
Can it be used commercially? Is the training protocol going to be open? Or just the weights released like Llama models?
Can it be used commercially? Is the training protocol going to be open? Or just the weights released like Llama models?
OpenAI has been good about sticking to the commercially-friendly MIT License for their OSS. (e.g. Whisper/tiktoken)
Researchers behave like alchemists when training these models, and the actions are not really reproducible.