Tag
The article highlights the heated debate over whether a new AI model is actually worse than before, pointing out that the real problem lies in the lack of reliable evaluation methods.