Is the Queen’s Indian Defence extinct at top levels?

Alpha Zero could beat Stockfish regardless of the opening (mostly). It has way more positional understanding than Stockfish, so it winning in a positional game is not all that surprising. Also, not every high-level game is decided in or by the opening. It’s still quite popular.

It's not quite as popular as in the 80's when Karpov was world champion, the trend is more for the Ragozin. Doubt AlapaZero has much to do with this though.

Not extinct. I think GMs just prefer the QGD.

AlphaZero is an atrocious chess engine that would have no chance against any serious chess engine, maybe even old Crafty would beat it. It has never won any chess game with equal hardware on both sides and it never could. 

AlphaZero also has zero understanding of chess due to the way it works. While standard chess engines have the input of human understanding the alphazero gimmick is that it has lack of human input, hence it has no understanding and is more trying to brute force it all while noticing patterns.  

Like most so-called "AI" Alphazero is pure gimmick nonsense and it's sad so many people fell for it.  

You know nothing about Alpha Zero. While the hardware debate is valid, the argument about it knowing nothing is not. Stockfish doesn’t understand patterns either, it just brute forces it. There are experts who say Alpha Zero has more human-like play, in the sense it looks at patterns. Do you even know how it got to it’s strength? Deep learning and neural networks, the same method our brain uses to learn and function. The brain is made out of wetware rather than hardware, but it still operates the same. By playing games against itself, Alpha Zero was able to learn how to play and what to value, not some input from a programmer. Alpha Zero doesn’t brute force problems, it learns. Stockfish does that. Please don’t make posts like this when you have no experience or knowledge in this field.

By saying it knows nothing about chess because it has no set values is irrational. While it is more like humans than Stockfish, it values things a bit differently. It’s better than us, so we may have something to learn about what it values. It creates those values based on patterns it sees give it an advantage.

The reason the Queens Indian defence is supposedly "extinct" on top levels(lol) is because it is known as a drawish opening. In those openings it gives equal chances and usually in those openings there is no way to break through for white without overextending his/her position, so in this case the game is about maintaining control but not taking any chances. So Queens Indian defense is actually much more similiar to its relative, the Kings Indian and the point is to gain equality, however the opening becomes good if white is aggressive and anti-positional with his/her play because if white is like that then they can lose more quickly than they can win.

Something to Note: The reasons many openings are not played in top-teir is not because they are inferior for the most part, it is due because they are drawish and since master's look to get wins, they avoid drawish openigs such as the Queens Indian, however if a Master wants a draw, or a relaxing positional and equal game, he will mostlikely do the Queens Indian and be comfortable within his position. 

This is why the Queens Indian is not played and that is why is not known in the top teir level, as it is drawish and because of this Master's don't have the chances they would have in otherwise an opening such as the Sicilain, Nimzo-Indian or Other defences, as since there is no imbalance and everything is in balance, equality is kept and the game doesn't go anywhere and is concluded a draw by an agreement by both Master's because both masters know they can't push for anything. And if they do push for anything they will have greater chances of losing than winning, so they draw the posiiton. 

So this is why the Queens Indian is not seen on top teir master play.

Beating an engine which is intentionally severely limited in strength, PLUS without an opening book (all QIDs played in that match are nothing new, but Stockfish did not "know"any theory and played it all by itself) is not an achievement.

The reason QID has declined in popularity a bit is that the QGD Ragozin and Vienna variations are currently extremely topical.

Well, let’s wait for the Chess.com version of those events. Lc0 vs. Stockfish 8 will be interesting.

QID is perfectly fine

it just has a drawish reputation at top level... like literally any defense 

I find nothing interesting in such a match, and Stockfish 8 is almost 4 years old.

SF 8 died a while ago 

Not really, Stockfish already won the latest fair computer engine contest for May to July 2020:  

https://en.wikipedia.org/wiki/TCEC_Season_18

Besides, LC0 has conventional chess engine processing as well which defeats the whole purpose of your argument. The neural net implementation may provide some advantages sometimes but it's not very good so far and if it was stockfish would also start implementing it as well. Nobody is saying neural network is a load of bs but the way it's been sold is a con.  

@SNUDOO

The reason I responded to you is because you said it was "perfectly fine".

To Clarify: It is perfectly fine if someone is ahead in a tourament(and doesn't want to play for a win) or wants to have a relaxing game before a tough challenger(because they don't want to use all of their energy on that "lesser challenger" and wants to save energy for the harder player.)

It is not a good opening to play all the time as you don't increase in rating and get higher by accepting draws, so I think it is good to be used when there is a specific reason but not just to play it. 

So it depends... 

@TeacherofPain

OP asked if QID was playable at top levels.

It is. I didn't take whether or not you wanted to win/lose in consideration

Playable? Any opening is playable, we are talking about the chances of an opening... 

Why do you think GrandMasters are selective of the different openings that can be played?(Depending if they are looking for a win or draw) Plus you always have to take the win/loss into consideration as you don't want to be on the losing the end of playing an opening that doesn't give you the chances you need to win, especially if you need a win in a tournament.
But again sometimes you don't need to go for a win in a tournament and can settle for a draw. So in that way playing for a draw would be fine.
So again it depends... But just know any opening is playable, heck 1.h4 is playable, it is not orthodoxed but it is playable as it is a move and who said that move isn't good if used in the right hands? So just to give you perspective of where I was coming from this is why the chances of an opening matters. As this is the reason why the QID is in the discussion in the first place. It is to see if it is an opening that is good(for a win or draw), not neccessarily playable, as any move in chess is playable.

I suspect that the Queen's Indian is just out of fashion at the moment.

Chess players are as fashion-conscious as teenage girls.

Just wait for a good novelty, then it will be all the rage!