The Future of Earth Observation Data, Right?

If the problem warrants hyperspectral every 16 days or something, then if that data exists, then yes, it can be used. But if the problem actually requires something that's, I don't know, five metre spatial resolution, then modus isn't really going to sort of do the right things. So historic data is very important but in modern machine learning and AI, this historic data will just be used as ingredient to create models. And those models are much easier to manage relatively than huge stores of past datasets. Otherwise, we'll keep on using raw ingredients to create new products and then we will grow them on BIS as quick.