Like Morphy and Fischer Alpha Zero is done.

Do not expect the DeepMind team to stay on chess for long. They will move on to physics or medicine. For all the gnashing of teeth all over the internet about "Was it a fair fight?" I think they could care less about chess. This was about the development of AI. The team will be on to bigger and better things. Maybe in 5 or 10 years AI will be developed enough to return to chess. We love our game but for them it was just a measurement tool.

Like Morphy and Fischer AlphaZero showed us something new, and now will leave us to ponder it's discoveries.

While it's a foolish assumption to believe that Deepmind will put any more effort in training its software on chess, I don't think that we saw the last about machine learning algorithms on chess. Actually it has just started a new generation of chess software, that will bring us tools to reach new levels of chess study. This year will be famous for chess software as the point of inflexion where chess Shanon B agorithms (selective search) eventually beat Shanon A algorithms (brute force search). This new algorithms like the one used by AlphaZero play much more humanwise than the current ones used by Stockfish, Komodo or The King. And it's strength is far easier to adjust to all levels of play and have the feel to be playing against another human. But its greatness is that its strength don't have a fixed value, it will improve over time as it learns about new branches. I'm sure that stockfish and komodo programers are right now implementing their own neural nets for a future version of their software IA. And as for AlphaZero, I don't think neither we saw the last of it on chess, as I'm sure Google will end up making it available to the public either as web app or mobile app, or both, and a lot of chess fans will use it to learn and play chess

Teach it Physics and Biology. I hope it's as good in those as it was in chess.

Since AlphaZero has not beaten any engines in an actual public match, I highly doubt they are planning to stop now.

They don't care about beating stockfish "officially". They don't care about the chess community. Or the game of chess. They are creating artificial intelligence. Our little egos and the game of chess mean nothing to them. Why continue? They have proved their point. Why put up with all the whining about what computer or what parameters were set how? Have you watched the videos analyzing the games? AlphaZero proved it can play at the highest levels. That's all they wanted.

Yes - next stop...a cure for cancer and a working nuclear fusion reactor! Hooray!

Right.

I'm sure the Candidate Master on the AlphaZero dev team has no interest whatsoever in seeing AlphaZero officially take the title.

If they don't do this officially this coming year, you can bet it's because AlphaZero can't actually beat Stockfish convincingly straight up yet.

P.S. If my lapdog follower Lawandorderking comes by, don't give him any more doggie treats.

DeepMind/Alpha Zero is owned by Google bro. You speak as if it's some dude who's now going to move onto something else. It's a division of Google. They beat the top GO players and kept on improving the engine. Maybe think its on its 7th generation now. Some members of he team will move on but they will replace them with others and keep developing.

Thanks, "bro".

I'm glad you know so much about Google.

I thought I knew a thing or two myself, having been at the Mt. View campus several times, knowing their director of Network Ops for decades, having a girlfriend who worked there, friends who work/worked there, etc.  I have a half dozen Google employee gel pens sitting in the cup next to me on my desk...

But clearly, you know much more than I do...what is your inside track?  Are you a software engineer at Google?  Are you on the DeepMind dev team?   That's amazing!  How do you know they are going to replace their CM?  That is some deep info...crap...are you managing the dev team?  That would explain everything...must be hard directing the team remotely from Kentucky, though.  Damn.  You must have mad skillz.

Lol. Although he does have a point. They stayed a while with Go, by playing Lee Sedol, and then Ke Jie a year later. Also they released a bunch of alphago vs alphago games for the go community to study them.

However, I'm not sure about his point of replacing their members, but he does have a point that they might stay with chess in the mean time.

I assumed Google bro was a division of Google for developing this thing. :"(

Time can be rented on Google's TPU Cloud so I would imagine that even if they step away from chess, someone will try to recreate the process and use that moving forward.

https://cloud.google.com/tpu/

I hope someone does. I was expressing what I expect to happen not what I want to happen. I hope they set something up soon. I'm not holding my breath.

If what they say about Alpha Zero creating opening novelties is true then GMs will be very interested. I'm certain that this will be developed.

Synthetic Chemistry is almost as much art as science: I wonder if AI will make much progress there, except for tweaking already known syntheses.

For example, I had to make some unusual orthoesters, an unusual class of compounds to begin with, and the procedure required an acid catalyst.  But my work failed and I finally realized it was the water that was part of the acids I was using that was ruining the reaction. Even the very tiny bit of water in even near-100% pure sulfuric acid was enough to mess it up.

Eventually, I thought of the resin pellets used to make hard water softer, nearly distilled: they are plastic, acidic, and have no water.  So I heated a bunch of resin pellets in an oven to drive off any last trace of water, dumped the pellets in the reaction flask, and in a couple of days I became the first person to ever lay eyes on some cyclic orthoesters as I distilled it, drop-by-drop, from the reaction vessel.

How do you get AI to think of that?

Great story! I am impressed.

Yo, dillweed...I'm on the other side of the argument and have been since I first posted.  It is the OP you are arguing with, apparently.  I don't think they are going to move on/shut down either.  Reading comprehension is your friend.

However...just because we're on the same side doesn't mean you understand much of anything.  Alpha Zero is not a consumer product.  The chess world/community will not get to buy it.  They will not monetize software that takes a supercomputer to run it, unless it is for corporations that can afford supercomputers.  That is, not the chess world.  They may do something else, they may create a "lite" chess engine someday, but no...you don't understand the world.  Maybe you should venture out of your flyover state and get to learning...

P.S. It's "an idiot". 

As I understand it, the reason DeepMind have been working on AIs to play games is that this is a task which has clarity, complexity, difficulty and has a good chance of providing enlightening insights that can be used in other areas. I agree with these conclusions to a greater extent than I agreed with IBM when they made a similar case for Deep Blue (there it was a lot more of a stretch to claim the technology was generalisable to other applications).