Go champion retires from professional play, cites undefeatable AI as the reason

Polycount

TS Evangelist
Staff member

That last set of victories is closely tied to the topic of today's news. South Korean Go champion Lee Se-dol, who was first beaten by AlphaGo back in 2016, has announced his decision to retire from professional play entirely, putting an end to a nearly 24-year-long career.

In a somber statement, Lee said the primary reason for his decision is the undefeatable nature of AI -- no matter how good a human player gets, he suggests, an AI will always top them eventually. "With the debut of AI in Go games, I've realized that I'm not at the top even if I become the number one through frantic efforts," he said in an interview with South Korean news agency Yonhap. "Even if I become the number one, there is an entity that cannot be defeated."

When Lee faced off against AlphaGo a couple years back, he only managed to score one win against the formidable AI. Still, that was an impressive feat in and of itself, and the fact that this victory came on the back of three consecutive losses speaks to Lee's resolve and skill. It would be easy for a lesser contender to simply give up, or play poorly due to frustration and low morale.

"Even if I become the number one, there is an entity that cannot be defeated."

There are certainly plenty of questions one could raise about the validity of an AI taking part in competitive gaming tournaments (digital or otherwise) against humans, and it's unfortunate to see a top-tier Go player retire due to this dilemma. Still, Lee seems to have come to terms with his new reality.

Moving forward, he says he will be focusing on spending time with his family and "resting." He will be playing one final celebratory match against the HanDol Go-playing AI next month, however.

Lee will be in good company following his departure from the Go scene -- AlphaGo itself retired from competitive play in 2017 after beating the world's top player (not Lee).

We wish Lee all the best, and we'll be interested to see the other effects DeepMind -- and AI in general -- might have on competitive gaming down the line.

Image credit: International Business Times

Permalink to story.

 

neeyik

TS Evangelist
Staff member
I think Lee's dilemma is entirely of his own making, or more rather, entirely in his own head. In any competitive sport, there are thousands of people all fighting against other competitors, or indeed themselves, trying to raise themselves to a better position/ranking/etc. If one was to give up, purely because there is always somebody out there who will beat you, then competitive sport would just cease to exist. The fact that a machine outplayed him at Go is beside the point. Do mathematicians give up because they can't find solutions to complex problems faster than a computer can? Clearly not.

I genuinely feel sorry for him, because it seems to me that he played Go less for the love of it, but more for just being better than anyone else out there. But I guess this is true for all top-of-the-field competitors...
 

psycros

TS Evangelist
I think Lee's dilemma is entirely of his own making, or more rather, entirely in his own head. In any competitive sport, there are thousands of people all fighting against other competitors, or indeed themselves, trying to raise themselves to a better position/ranking/etc. If one was to give up, purely because there is always somebody out there who will beat you, then competitive sport would just cease to exist. The fact that a machine outplayed him at Go is beside the point. Do mathematicians give up because they can't find solutions to complex problems faster than a computer can? Clearly not.

I genuinely feel sorry for him, because it seems to me that he played Go less for the love of it, but more for just being better than anyone else out there. But I guess this is true for all top-of-the-field competitors...
Apples and oranges - computers don't COMPETE against mathematicians but aid them in their endeavors. AI and robots shouldn't be competing against humans in any field except games where an AI enemy is used to simulate a real one, and in that situation the measure of a good AI is how human-like its behavior appears. If this guy is going to be continually forced to compete against machines with flawless memory and infinitely scalable parallel processing then why bother? Imagine if they started putting wheeled robots in foot races - nobody would tolerate that. How anyone could fail to comprehend such a simple situation is beyond me.
 

neeyik

TS Evangelist
Staff member
If this guy is going to be continually forced to compete against machines with flawless memory and infinitely scalable parallel processing then why bother?
He's not; nobody in the competitive Go circuit is. Lee can easily stay in competition and reach #1 status again - or more rather, he can't at the moment because he quit the association that handles such entries 3 years ago.

computers don't COMPETE against mathematicians but aid them in their endeavors.
For the most part, this is certainly true. However, competitive mathematics does exist (e.g. the IMO) and just like in international Go tournaments, the use of computers of any kind is prohibited. However, there are attempts (one such example) to pitch AI against the human competitors in the same field. It will be interesting to see the outcome of such attempts.
 
  • Like
Reactions: Polycount

amghwk

TS Guru
He should keep playing with the AI to improve and keep his skills polished, AND continue tournaments against Human players. Rather than retiring. But to each his own.