WTF did I just watch?

[CrusaderFrank]

First, we need an AI forum; it’s the future and it’s already here.

I just watched an annotated game between Magnus Carlson and Stockfish, where Stockfish game World Champ Carlson knight odds - Stockfish played without his Queenside knight and handily demolished the human champ. THE VIDEO WAS ANNOTATED BY STOCKFISH. The commentary was insightful and downright funny at times “…because I am Stockfish”

Magnus is rated at over 2800, Stockfish just over 4000!

When I figure out how to link from the phone I’ll post the video

 

[luiza]

Gave up the game for Bridge many years ago but still follow it in "broad stroke " terms .

4000 is so far ahead that understandably I wonder if you have been "joked" .

My limited understanding of AI is that its results cannot improve on the potential/ capabilities of its programmer(s)
.
Which at best comes from a very savvy group of IMs .

If I am even near right , this story is a Xmas Cracker joke .

 

[CrusaderFrank]

Gave up the game for Bridge many years ago but still follow it in "broad stroke " terms .

4000 is so far ahead that understandably I wonder if you have been "joked" .

My limited understanding of AI is that its results cannot improve on the potential/ capabilities of its programmer(s)
.
Which at best comes from a very savvy group of IMs .

If I am even near right , this story is a Xmas Cracker joke .

Stockfish has been around for a while and it’s no joke. A video last week showed it finding a flaw in a chess problem solution that had been accepted since the 1950’s

 

[luiza]

Keep me posted .

I have come from completely disbelieving to dubious .

Although there is no translation of a FIDE rating to IQ , my feeling is that it might equate to something like 300 which ,

if true , could be very worrying .

Suddenly our species would be redundant , especially if some sort of consciousness emerged .

They would detest us and cancel us out .

 

[JoeMoma]

Gave up the game for Bridge many years ago but still follow it in "broad stroke " terms .

4000 is so far ahead that understandably I wonder if you have been "joked" .

My limited understanding of AI is that its results cannot improve on the potential/ capabilities of its programmer(s)
.
Which at best comes from a very savvy group of IMs .

If I am even near right , this story is a Xmas Cracker joke .

AI can "teach its self" chess strategy by running more chess games against itself than a human chess master can play in a million lifetimes. As such, the patterns of moves (strategy) "learned" by the AI can greatly surpass even the greatest of human grand masters. AI isn't programmed in the traditional sense , in other words, humans are directly coding the chess strategy used by humans into the AI.

 

[luiza]

I do not understand the statement , " isn't programmed in the traditional sense"..

If AI aims to deal with every move sequence possible -- an outlandish number---it seems that this is precisely what we
as a species already try to do , except we are physically incapable of achieving it : because the machine can travel further -- handle more permutations -- and one day will presumably be able to do all of them at the speed of light .

If so , that is only improving on our present and traditional programming , just doing more calculations and all at a much faster speed .Therefore , the traditional sense has been maintained but with improvements .

I imagine you know better, so I would like to know, why, and, how .

 

[CrusaderFrank]