Re: And AlphaGo Zero
I would not call it "already fairly big in AI." In terms of number of dollars invested in AI, yes they are fairly big. Way more invested save the other large players. There's Watson, of course, and TrueNorth, their co-invention Systems of Neuromorphic Adaptive Plastic Scalable Electronics, or SyNAPSE. Their problem is that they lack customers, paying customers, one's paying huge amounts of cash for access on a continuing basis in particular verticals. As I recall, their attempt to break into and branch out in the medical vertical died when it failed to deliver fast, accurate cancer detections. [Not going to hunt down that referrence.]
They really do have a chance if they can keep Ginni Rometty from even worse redundancies in that department, not likely. Everything must go to meet share price goals of her own invention. And it doesn't help when the IBM Cloud keeps dying due to clerical gaffes a near novice in the web hosting business would manage to avoid.
I love IBM. The old one that put me on the path of database, system, software, and hardware engineering. For a start. I can still picture pages from the man pages in my mind. TrueNorth is very, very interesting due to the 70 mW power profile with serious performance. Just, Ginni needs to go so what few engineers left can get back to their real job. New inventions with what used to be the best documentation in the business.