> 1. winter scenario<p>> 2. holy shit scenario<p>3. continued enthusiasm, based on actual results<p>There are so many applications of current deep learning techniques that we have decades of work ahead, and that will keep AI from wintering out. If we applied what we know so far on current problems, in various domains, it would be revolutionary by itself.<p>The state of AI penetration into the world is like personal computing in 1980's. The really great applications haven't yet been realized, and it's just a matter of applying knowledge.