Human Go Champion Backtracks On Vow To Never Face An AI Opponent Again (engadget.com)

Back in May, Google's AlphaGo AI defeated the human world champion Ke Jie in a three-part match. After it was over, Jie vowed never to play a computer again. But apparently something has changed his mind because Chinese news sources report that Jie will once again play an artificial intelligence at an AI tournament to be held in China in April 2018. Engadget reports: Ke Jie is one of the tournament's ambassadors, and he will play against the AI Tianrang. Normally, a human representative places pieces on behalf of the AI, but in this case, a robotic arm developed by Fuzhou University will fulfill that role. Tianrang previously ascended to the semi-finals of Japan's AI Go tournament, called AI Ryusei, earlier this month. Tencent's AI was the ultimate winner of that tournament. The complement of AI competitors for the Chinese tournament are Tianrang (Shanghai), DeepZenGo (Japan), CGI (Taipai) and more. Google DeepMind's AlphaGo has since retired from competition, so it will not be playing in the tournament.

Comments

Why he changed his mind

[Anonymous Coward]

Most likely, this is why he changed his mind: $.

Re:Why he changed his mind

[HumanWiki]

Could also have been strongly suggested that he play again if he knows what's good for him... Wouldn't surprise me in the least.

Re:Why he changed his mind

[hcs_$reboot]

Maybe because Google's AlphaGo (that defeated Ke Jie) retired from competition?

Re:Why he changed his mind

[ortholattice]

Or possibly he's been studying the AlphaGo strategies that beat him and thinks he has found a weakness.

Re:Why he changed his mind

[ayesnymous]

They would actually threaten a guy over Go?

WTF??

[mark-t]

You don't "back out" of a vow... you break it.

As if breaking vows isn't considered serious enough, the fact that one would consider that a vow could simply be "backed out on" like it never happened in the first place really shows a very large misunderstanding of what it means to even make a vow in the first place.

Re:WTF??

[ShanghaiBill]

The word "vow" was made up by the journalist writing TFA. Ke Jie never used that word. His statement was more in the form of an off-hand remark in a blog post, and was written in Chinese. You are complaining about the nuance of an English word, when neither English nor that particular word was actually used.

Re:WTF??

[mark-t]

One would ordinarily think that somebody who was worthy of the title "journalist" would know the definitions of words well enough to realize that "backing out of a vow" is a complete and utter contradiction in terms.

Re:WTF??

[mark-t]

I was not meaning to imply anything about the morality of breaking vows, only that when one changes their mind about something they may have vowed to do, then they are, in fact, actually breaking a vow. The only reason to even call it "backing out" is to perhaps maybe make it sound like it isn't really that bad, or perhaps more specifically to rationalize how it's not that serious because actually "breaking" a vow or a promise sounds a whole lot more serious and feels easier to judge harshly against than just trying to act like it never happened in the first place.

Re:WTF??

[ShanghaiBill]

One would ordinarily think that somebody who was worthy of the title "journalist" ...

Here is a complete and comprehensive list of all the qualifications, skills, and credentials needed to acquire the esteemed and august title of "Journalist": { }

Re:WTF??

[religionofpeas]

One would ordinarily think that somebody who was worthy of the title "journalist" would know the definitions of words

I recommend not making too many assumptions in life, especially not where it concerns the expertise of journalists.

Re:Human, shmuman!

[ShanghaiBill]

At this point, would it not make more sense for AlphaGo to come out of retirement to play Tianrang for the title?

No. Software doesn't "come out of retirement". To have a chance in the tournament, AlphaGo (actually AlphaGo-Zero) would have to be trained with thousands of GPU cores for thousands of hours. It would likely also need both wider and deeper layers, and some of Google's best AI researchers devoted to tuning it. That is a lot of resources. For what? If Google wins, nobody cares, because AlphaGo already beat the human champion, and plenty of other programs. But if AlphaGo loses, it will be big news and Tencent (the likely winner) will get plenty of good PR and perhaps more leverage to hire talent.

With nothing to win, and plenty to lose, Google is smart to let AlphaGo "stay in retirement".

Re:Human, shmuman!

[rodrigoandrade]

But sir. Nobody worries about upsetting a droid.

Re:Funny you should say that...

[ShanghaiBill]

I don't think people have really been taking the full seriousness of the word 'vow'

You make it sound like that is a bad thing. Societies that place an importance on "honor" tend to be violent, intolerant, and economically backwards. It is better to be open minded, willing to compromise, and less concerned about what other people think of you.

Re:Funny you should say that...

[mark-t]

I wasn't addressing the matter of how flippantly some people may or may not make vows.... I was addressing that a vow is not something that can be backtracked on liike it was never made. You can break a vow, but you can't just go back on one, any more than you can unshatter a piece of glass that you hit with a hammer.

Re:Funny you should say that...

[religionofpeas]

Was it really a vow, though ? Or did he just state his intentions ?

Re:Funny you should say that...

[mark-t]

That's sort of my point... calling something a vow when it's not one is like calling red black or a duck a cat. It's entirely the wrong word to use. Vows can, as I said, most certainly be broken, but you can't undo one without literally reversing the entire flow of time.