 So now what we just saw is what happens if we only have one axis that is interesting. But what if there's more dimensions? There was a recent nice paper that showed how we can deal with such system. What can we do? We may expect to separate the different dimensions using methods like dimensionality reduction. And what do we expect? That early on the dimensions will be pretty independent of one another. And if we have growth that is very rapid, then if you want the dimension that first makes it will become dominating very quickly. It will go like this and then will come the second dimension, then will come the third dimension and so forth. Maybe towards the end there will be some interaction where you have then big components with one big component taking some weight from another big component. But at least in the early parts of that we expect everything to start and then for them to kick in one after the other. So why don't we explore that in a big annual network?