The AlphaGo Milestone

My friend is scared.  We’ve ducked in to a basement Starbucks.  How many Starbucks are there these days in Beijing? I’ve been in the hotel next to this mall many times over the years, but I’ve never wandered down here.  Other people have.  The place is packed.  China, never at a loss for people. 

We were supposed to discuss a client we’re both working with.  My concerns are pretty simple:  how much can we make this year?  But as I said he was frightened.   The Google AlphaGo application had beat the Korean Go master, Mr. Lee Se-dol, three games in a row.  Mr. Lee was completely dumfounded.  “I am very surprised because I never thought I would loose.”  In what seems like predictable hubris he went into the game cocky.  The machine hadn’t played anyone with real skill yet, he maintained.  But this was about to change.   

Any my friend found all this rather disturbing.  “Now it’s only a matter of time until we are all simply cultivated as pets.  Pets, or slaves.”  In his mind, automation would expand rapidly now.  Machines would do all manual jobs better than humans and humans.  A cadre of the “free” in concert with smart machines would placate and manage the plebs.  The AlphaGo milestone was a harbinger of humanity’s pending redundancy. 

The program, to-date, is only mimicking human intuition.  Learning as it goes, it imitates.  But how much longer do we have to wait for the imitation of moral choices, adherence to ethical boundaries that are also imitated so faithfully that they feel like a machine with free choice?  Swiftly, this will no longer be something that the US, or the western world, can drive the agenda on.  Our own devil-you-know moral framework, where legislation catches up with innovation in a clumsy but reassuring way, after a few years lag time, is irrelevant here in China.  A different ethical calculus with fewer checks and balances will yield different innovation and different moral frontiers confronted. 

I considered the Starbucks.  Would I pay extra for the boutique experience of buying coffee from a real person, instead of machine?  I might, but who cares?  Most people wouldn’t.  All of this, driving cars, engaging wait staff, bothering to meet in person, soon to be so quaint and anachronistic. 

And so, clutching straws certainly, I was glad to see that we could still strike a blow for human intuition and for Korean hubris.  Mr. Lee managed to kick AlphaGo’s ass in the fourth game.  He suggested that he made a “surprise” move that made the program act as if it had a bug.  He was also convinced that AlphaGo did better when it held the white stone.  Hard to imagine how one could “surprise” a program, or outperform based on color, but hey, to the winner goes the spoils.  Surely this drama though will soon appear as quaint and otherworldly as John Henry’s battle to the finish with a steam powered hammer.  Brain and brawn are only venerated until something else can surpass them.