The Importance of Collaboration in Machine Learning

It's very, very hard to find a problem that matters where current tools aren't good enough to do something very useful. And if you can find that, you're like golden on research. Sometimes what happens is that it doesn't require particularly new machine learning, but it might get you thinking in very interesting directions. I'm curious if we could go and do like a case study of one of your pieces of work on CRISPR guide design.