Clusters have two dimensions. One is very geographic, because we know that where we have a certain number of players, we build out from there, and it's really important for that ecosystem, but they are also nationally focused.
We are exceedingly small in Canada. I think Jim Hinton provided evidence a couple of days ago about AI. Our patents are spread out across Canada. There is no one place. If we just focused on the Vector Institute in Toronto, we would lose. It has to be pan-Canadian.
Somehow, we have to bring the patents that exist across Canada and make them available to Canadian firms. From that perspective, we can't have firms outside. I think that's the idea behind superclusters. It's not to do original research; it's to build the infrastructure that allows that sharing and common base.
We want competition within Canada, kind of like intramural sports, but when we go on the outside we want Canada's A-team, so we have to bring it together. In my mind, that's what the supercluster is about.
That doesn't negate the importance of local clusters, such as the ones you're talking about or the MNI, where we build communities that involve universities and civil society. The cheapest way to deliver benefit in mental health may not be through pharmaceutical products; it may be through social services. We have to bring together all of these actors locally to develop new solutions. To me, that's not the supercluster; that's more traditional clusters.