Is There a Future for Machine Learning?

We're starting to see applications involving modeling. So in that case, imagine a scenario where you have one party that has a large list of users and only a subset of those users are actually identified via email. And you might be an advertiser and then we'll do a match. If I'm that party with the large unidentified user base, what I'm interested in actually doing is using the list of overlapping users to then train a model That will score all of my other users in terms of how similar they are to the overlap.