Will computers ever solve chess?

Not for sure...engines run using a set of valuations based off of human understanding of chess.  Until engine developers bootstrap the engine's own results over years of play without using human opening books or position valuations to come up with a set of valuations built from the ground up without human "seeding", all that engines can prove is that they (on average) know how to play way better than humans and equal to themselves.  That does not mean chess is then "solved".

I think strength and weakness of specific openings might become more clear when played by two super-computers. For example, an opening like a3 is considered downright silly. But, is it really that bad or not? That can be known if two supercomputers play it and by comparing the results of this opening to other openings.

Here is what I know about solving chess, give it a read, its very informative: http://pintient.com/2ypL

I'm at a lose here. How can 1. a3 be "silly" ??? It must be considered as any other move for White. If you want to say White gave away 1st move initiative, fine. Black plays e5 or d5 and the game is evaluated as Black being slightly better. but it's White's move. Computers evaluate material. The test is, is it possible to evaluate outside the programming, can it learn independently.

Search DeepMind and the AlphaGo program. It played opening moves not ever played by humans. Not considered to be the "best". Yet the program convincingly bested the world's best. Had it achieved some kind of AI?

Yes, that is the direction I was talking about.  Having an engine play human opening book moves and use human-derived piece valuations, etc. limits the engine's top end play.  It won't always be that way, though, but until there's a significant change in approach, and engines are playing at a level where even Carlsen can't begin to analyze an engine game and pretend to understand why it chose X over Y, engines cannot be said to be playing up to their full potential and chess certainly can't be "solved".

Exactly.

 AlphaGo proved alternative opening choices, not previously considered by humans, proved to be effective. The interesting part, the opening choices were not pre-programmed, but played by AlphaGo based on past learning. The ideas are now being copied by top Go players, expanding their understanding of the game. Unfortunately, the developers have "retired" the program from competition, as they wish to move on to other fields.

Apparently, there is a group getting ready to try to solve some game more complicated than checkers by volunteer distributed computing. It might be a version of Go or chess. If anyone is interested, they might be able to contact the group at (this website).

Their last project was sudoku (9 x 9 game) but I think they are thinking of something bigger now.

There is no point in trying to solve chess using the present computing resources of our planet. We need a lot more.

Bare minimum is the rules of chess (computers could even learn these, by example or by reinforcement learning!) Given the rules of chess, a computer can see the results of games. They can also see what is possible more than one move ahead. Current chess programs are mostly hard coded, but they do use machine learning in some cases, and AlphaGo uses it in a huge way.

Educated people have been pointing this out for the past 1300 posts.

Why can't readers get with the program and have there own eureka moment regarding same ?? 

I guess we are all "Brain Hacked" by the ecosystem, as per the 60 Minutes show last night on CBS.  Check it out, on streaming, or whatever ecosystem you might frequent.

Please refrain from making predictions based on 'other stuff was deemed impossible but they were wrong, this is deemed impossible, therefore this is possible.'

At the time predictions of future computer power were hard to make. Currently we have a pretty accurate view of the future and can safely asses that even if we are off by several orders of magnitude (in layman terms: a fuckity fuck fuckton) it's still not even close to possible.

You make it seem like "real innovation" is magic, let me tell you a secret:

....

*it isn't*

Engineers are still bound to natural laws. No matter how much you say "this and that was once impossible ... but now it's here, therefore this impossible thing is also possible", it still doesn't change the figures.

If you want to get an idea how certain this is:
- look up how many different games can be played (at the very least 10^120 aka Shannon number and more likely around 10^123)

- look up how many atoms are present in the observable universe (4*10^79 ~ 4*10^81)

- draw a conclusion

You are literally proposing to have a computer do more work than counting all the atoms in the universe at the very least 10^29 times... That is, recounting every. single. atom.
100000000000000000000000000000 times over.

Now you can say "maybe in the future", "maybe with more technology", "Maybe our engineers are just too stupid"... Orrr you could have a look at reality and say "yeah.. that'll never happen"

Not correct. The entire game-tree of chess does not have to be searched to find a forced mate.

How about you prove me wrong then . I bet that with the new quantum computers you could do it over the weekend!

Your sequence of logic was already proven wrong.

Good thing you cannot explain why.

computers arent capable of solving chess.  they can only do what they are programmed to do by a human.  if a human finds a mathematical formula then the computer will be able to process it. but the human equation will always come first.  we might find a mathematical forumula but i seriously doubt it because it isnt like counting cards in black jack.