

Li-ion technology has huge factories behind it, so economies of scale apply here. The first Na-ion battery factories have just started, so everything is more expensive to manufacture on a small scale. However, the ingredients are cheaper and easily available. Once they ramp up production, we can make a fair comparison between the two.
I have a feeling LIBs are going to be more expensive, but they won’t disappear since high energy density is very handy in mobile applications like cars and phones. NIBs are probably going to end up being a lot cheaper, which should make them a popular option in all the less demanding applications, like grid energy storage, kitchen scales, and anything in between.












Well, they are language models after all. They have data on language, not real life. When you go beyond language as a training data, you can expect better results. In the meantime, these kinds of problems aren’t going anywhere.