Lee Sedol, the only human to beat AlphaGo retires

South Korean Go master Lee Se-dol submitted his letter of retirement to Korea Baduk Association and retired on Tuesday. Notably, In 2016, he surprised the world with the match between AlphaGo of Google's DeepMind Technologies. Watch this video to findout more

#Leesedol #Baduk #Alphago #Leesedolretirement

Related article:

[mr_art]

He brought honor to humanity, a very sharp player with an innate sense for stability in his play games. Calm and modern playing style. His move 78 is proof that there can always be flaws in perfect systems. Ultimately I am sure he had his reasons, but maybe having AI play this game was not the right ethical thing. So many people left KGS and IGS. I guess Hikaru no GO needs a sequel with AI as the opponent. And the protagonist needs to be a 9d going for 10 honorary dan hahahaha

[scivolanto]

He's not the only one to have beaten an AI player, but the only one to have beaten AlphaGo. There are a lot of bad AI players anyone can beat. Anyway, he is a living legend.

[mandor]

I do not understand why. Magnus Carlsen doesn't retire althouth there is LeelaZero ... Chess, Go, they are also sports, .. the cars run faster that any athletes, but athletes are still running indeed. And Lee Sedol in fact wins over this monster one game (It was literally amazing to saw this intuition).

[Kitora_Su]

The title is a bit misleading. Alpha Go version 18 won by 4-1 against Lee Sedol. That was the version where Alpha Go was far from perfect and the team didn't have enough time to fix the issues. (Everything was mentioned about the issues in the Alpha Go movie/documentary by Deepmind). Later the perfect version of AlphaGo beaten every player in 60 games consecutively (60-0 in AlphaGo's favour). But, then AlphaZero that learned everything from scratch and never had any human input i.e., never learned from human examples and played against itself had beaten AlphaGO Master by 100-0 in AlphaZero's favour.

Edit- It only needed one system's computational Power since it didn't use billions of data/examples. The only data it had were the Rules of Go. Purely amazing Computer Science and Ai engineering.

[yixing]

2020, Engineers compete in GO with their trained models

[keselekbakiak]

Lee see dol quit because the involvement of politics in go

[diamondmeeple]

It was considered too hard for an AI to become the best in Go... so he thought...

He wanted to improve and develop his skill .... but then all of the magic of discovery got killed by the superior new AI, and humans were turned into copycats and memory sticks.

[mr_art]

Came back to this video to say I hate AlphaGo and Deepmind. :) They honestly did a low jab to the go community. Less and less people are showing up and I can't find some of my digital pals. I wish AlphaGo was never created, but that's just wishful thinking.

[Hammett]

I expect this sort of game to be popular in hell.