Just watched this talk. Definitely worth watching.

It's cool to see how Hinton worked on solving a ML problem and ended up gaining insight on two problems in biology that seem unrelated to each other, but are perfectly related by his ML solution.